This report aims to provide a more comprehensive picture of the full health loss attributable to diabetes and chronic Diabetes and chronic kidney disease kidney disease (CKD). It quantifies the impact of diabetes and CKD on the burden of other diseases for which there as risks for other diseases is evidence of a causal association (‘linked diseases’) to estimate the indirect burden caused by these 2 diseases. It uses disease burden estimates from the Australian Burden of Disease Study 2011 and extends the standard Australian Burden of Disease Study 2011 approach for analysis of risk factors to model diabetes and CKD as risk factors. When the indirect burden due to linked diseases was taken into account, the collective burden due to diabetes was 1.9 times as high, and CKD was 2.1 times as high, as their direct burden.
AUSTRALIAN BURDEN OF DISEASE STUDY SERIES NO. 8
AUSTRALIAN BURDEN OF DISEASE STUDY SERIES Number 8
Diabetes and chronic kidney disease as risks for other diseases
Australian Burden of Disease Study 2011
Australian Institute of Health and Welfare Canberra Cat. no. BOD 9
The Australian Institute of Health and Welfare is a major national agency that provides reliable, regular and relevant information and statistics on Australia’s health and welfare. The Institute’s purpose is to provide authoritative information and statistics to promote better health and wellbeing among Australians.
© Australian Institute of Health and Welfare 2016
This product, excluding the AIHW logo, Commonwealth Coat of Arms and any material owned by a third party or protected by a trademark, has been released under a Creative Commons BY 3.0 (CC-BY 3.0) licence. Excluded material owned by third parties may include, for example, design and layout, images obtained under licence from third parties and signatures. We have made all reasonable efforts to identify and label material owned by third parties.
You may distribute, remix and build upon this work. However, you must attribute the AIHW as the copyright holder of the work in compliance with our attribution policy available at
This publication is part of the Australian Institute of Health and Welfare’s Australian Burden of Disease Study series. A complete list of the Institute’s publications is available from the Institute’s website
ISSN 2204-4108 (PDF) ISSN 2006-4508 (Print) ISBN 978-1-76054-045-6 (PDF) ISBN 978-1-76054-047-0 (Print)
Suggested citation Australian Institute of Health and Welfare 2016. Diabetes and chronic kidney disease as risks for other diseases. Australian Burden of Disease Study 2011. Australian Burden of Disease Study series no. 8. Cat. no. BOD 9. Canberra: AIHW.
Australian Institute of Health and Welfare Director Mr Barry Sandison
Any enquiries relating to copyright or comments on this publication should be directed to: Digital and Media Communications Unit Australian Institute of Health and Welfare GPO Box 570 Canberra ACT 2601 Tel: (02) 6244 1000 Email: [email protected]
Published by the Australian Institute of Health and Welfare
This publication is printed in accordance with ISO 14001 (Environmental Management Systems) and ISO 9001 (Quality Management Systems). The paper is sourced from sustainably managed certified forests.
Please note that there is the potential for minor revisions of data in this report. Please check the online version at
Contents
Acknowledgments...... iv Abbreviations ...... v Symbols ...... vi Summary ...... vi 1 Introduction ...... 1 2 Methods ...... 8 3 Diabetes estimates ...... 14 3.1 Diabetes burden in the ABDS 2011 ...... 14 3.2 What is the collective burden of diabetes? ...... 15 3.3 How much burden from linked diseases is due to diabetes? ...... 16 3.4 Indirect diabetes burden by linked disease ...... 19 3.5 Estimated diabetes burden in 2020 under 2 different scenarios ...... 28 3.6 Scenario differences by age ...... 30 4 Chronic kidney disease estimates ...... 31 4.1 Chronic kidney disease burden in the ABDS 2011 ...... 31 4.2 What is the collective burden of chronic kidney disease? ...... 32 4.3 How much burden from other diseases is due to chronic kidney disease? ...... 33 4.4 Indirect CKD burden by linked disease ...... 35 4.5 Estimated CKD burden in 2020 under 2 different scenarios ...... 39 4.6 Scenario differences by age ...... 41 5 Discussion ...... 42 Appendix A: Selection of effect sizes for diabetes ...... 47 Appendix B: Selection of effect sizes for chronic kidney disease ...... 56 Appendix C: Methods for scenario modelling ...... 60 Appendix D: Data sources ...... 69 Glossary ...... 72 References ...... 74 List of tables ...... 81 List of figures ...... 82 List of boxes ...... 84 Related publications ...... 85
iii
Acknowledgments
This report was prepared by staff in the Australian Burden of Disease Unit and the Cardiovascular, Diabetes and Kidney Unit of the Australian Institute of Health and Welfare (AIHW). The main authors were Wendy Ho, Melanie Dunford and Vanessa Prescott, under the guidance of Michelle Gourley, Sushma Mathur and Lynelle Moon. Other AIHW staff who made a substantial contribution were Michael DeLooper and Roslyn Seselja. Melissa Goodwin and David Whitelaw from the AIHW provided valuable input and statistical advice, while Geoff Neideck, Fadwa Al-Yaman and Justin Harvey provided constructive comments and review. Their contributions are gratefully acknowledged. The authors would also like to acknowledge the comments received from the Epidemiology Branch of the Public Health Division from the Western Australian Department of Health. The report was prepared under the guidance of the National Vascular Diseases Monitoring Advisory Group (NVDMAG). Members are: Erin Lalor (Chair), Elizabeth Flynn and Bernie Towler. The NVDMAG includes the following expert advisory groups and their members: Cardiovascular Disease Expert Advisory Group: Andrew Tonkin (Chair), Tom Briffa, Derek Chew, Annette Dobson, Mandy Thrift and Mark Nelson Diabetes Expert Advisory Group: Jonathan Shaw (Chair), Stephen Colagiuri, Maria Craig, Wendy Davis, Mark Harris, Greg Johnson, Glynis Ross and Sophia Zoungas Chronic Kidney Disease Expert Advisory Group: Steven Chadban (Chair), Alan Cass, Jeremy Chapman, Joan Cunningham, Bettina Douglas, Stephen McDonald and David Parker. The Australian Government Department of Health funded this report. The authors acknowledge the valuable comments from individual staff members.
iv
Abbreviations
AB attributable burden
ABDS Australian Burden of Disease Study
ABS Australian Bureau of Statistics
AHS Australian Health Survey
AIHW Australian Institute of Health and Welfare
ANZDATA the Australia and New Zealand Dialysis and Transplant Registry
CHD coronary heart disease
CI confidence interval
CKD chronic kidney disease
DALY disability-adjusted life years e.g. for example eGFR estimated glomerular filtration rate
ESKD end-stage kidney disease
GBD Global Burden of Disease (study)
HR hazard ratio
OR odds ratio
PAF population attributable fraction
RR relative risk
WHO World Health Organization
YLD years lived with disability
YLL years of life lost
v
Symbols mg/dL milligrams per decilitre mmol/L millimoles per litre mL/min millilitres per minute
1.73m2 a standard body surface area (measure for eGFR) > greater than ≥ greater than or equal to < less than ≤ less than or equal to % per cent — nil or rounded to zero + plus – minus ± plus minus
vi
Summary Diabetes (type 1 and type 2) and chronic kidney disease (CKD) can act as risk factors for other diseases such as coronary heart disease, stroke and dementia. The Australian Burden of Disease Study (ABDS) 2011 only reported on the direct burden of diseases. To fully account for the health loss attributable to a specific disease, a diseases-as-risks approach (linked disease) can be used to estimate their ‘indirect’ or additional burden. The direct and indirect burden can be added to estimate their collective burden. Diabetes and CKD burden doubled when taking into account indirect burden The ABDS 2011 reported that diabetes and CKD were responsible for 2.3% and 0.9% respectively of the total burden of disease and injury in Australia in 2011 (the direct burden). When the indirect burden due to linked diseases was taken into account: • the collective burden due to diabetes was 1.9 times as high, and CKD was 2.1 times as high, as their direct burden • the indirect diabetes burden varied by sex, with males experiencing 31% more burden than females • the indirect burden due to diabetes and CKD occurred at a later age than direct burden, being responsible for over 50% of the collective diabetes burden and 65% of the collective CKD burden from age 75 onwards. Of the 12 linked diseases examined for diabetes: • the burden attributable to diabetes was highest for coronary heart disease, stroke and CKD—together accounting for 75% of the indirect diabetes burden measured • diabetes was responsible for 21% of the CKD burden, 14% of the stroke burden, 12% of the liver cancer burden and 11% of the coronary heart disease burden. Of the 4 linked diseases examined for CKD: • the burden attributable to CKD was highest for coronary heart disease—accounting for almost half (48%) of the indirect CKD burden measured • CKD was responsible for 19% of peripheral vascular disease burden, 8% of dementia burden and 7% of stroke burden. Around one-fifth of future diabetes burden could be avoided if the current rise in diabetes is halted If the current trends in diabetes and CKD prevalence and mortality continued to 2020, the estimated collective diabetes burden is projected to be 1.6 times as high as in 2011, and the estimated collective CKD burden is projected to be 1.4 times as high. This compares to rate ratios of 1.3 if prevalence and mortality rates are maintained at 2011 levels to 2020, (which reflects population growth and ageing). Put differently, if the current rise in these diseases is halted, 21% of future diabetes burden and 5% of future CKD burden could be avoided. In terms of indirect burden, the greatest gains are expected to be made in those aged 65–94 for diabetes, and 65–84 for CKD, where around 36% of diabetes attributable burden and 15% of CKD attributable burden could be avoided by 2020 if the prevalence of these diseases is maintained at 2011 levels. Results from this study could be used to inform population health monitoring and may assist in the development of chronic disease policy.
vii
viii
1 Introduction
Diabetes mellitus (diabetes) and chronic kidney disease (CKD) are increasingly common in Australia, partly related to changing lifestyle factors and ageing population factors. These chronic diseases have major impacts on health and other support services. It is well known that they have common modifiable risk factors—such as smoking, obesity, physical inactivity and poor nutrition—that can often be prevented or effectively controlled through behavioural change, medications and health-care interventions (AIHW 2015). Some diseases can also act as risk factors for other diseases (referred to as ‘linked’ diseases in this report). Evidence suggests people with diabetes or CKD are at higher risk for other chronic diseases (Liao et al. 2015; Tonelli et al. 2012). For example, diabetes can increase the risk of coronary heart disease, stroke, peripheral vascular disease, CKD, certain cancers and dementia (Cheng et al. 2014; Emerging Risk Factors Collaboration 2010; Prince et al. 2014; Vigneri et al. 2009). CKD is associated with the development of cardiovascular diseases such as coronary heart disease, stroke and peripheral vascular disease (Tong & Stevenson 2007), as well as dementia. The burden of each disease as calculated in the Australian Burden of Disease Study (ABDS) 2011 only captured the ‘direct’ disease burden. For example, the direct burden of diabetes is the burden directly related to diabetes, such as diagnosed diabetes, diabetic neuropathy and retinopathy. To fully account for the health loss attributable to diabetes, the ‘indirect’ or additional burden from diseases linked to diabetes must also be calculated, including the proportion of burden of linked diseases caused by diabetes complications (for example coronary heart disease and stroke). The direct and indirect burden are added together to estimate the collective burden (Figure 1.1). To date, the indirect burden of diabetes and CKD is unknown in Australia.
Direct Indirect Collective burden burden burden
Examples of direct burden Examples of indirect burden • Diagnosed diabetes • Coronary heart disease • Diabetic neuropathy • Stroke • Diabetic foot ulcer • Peripheral vascular disease • Amputation due to diabetes • Dementia • Vision impairment due to • Chronic kidney disease diabetes • Some cancers Figure 1.1: Examples of direct and indirect burden of diabetes
Diabetes and chronic kidney disease as risks for other diseases 1
1.1 This report, its aims and analytical approach This study aims to provide a more comprehensive picture of the full health loss attributable to diabetes and CKD. It quantifies the impact of diabetes and CKD on the additional burden of linked diseases as they act as risk factors. This effect is termed ‘disease-as-risks’, and presents estimates of the collective burden of diabetes and CKD in Australia in 2011. It uses estimates of disease burden from the ABDS 2011 (AIHW 2016a) and extends the standard approach for analysis of risk factors for diseases to assess the impact of diabetes and CKD as risk factors for other diseases. The aims of this report are to: • assess the indirect disease burden (DALY) attributable to diabetes and CKD • determine the collective burden (indirect and direct) of diabetes and CKD • explore the effect of reducing diabetes or CKD prevalence on the burden of various linked diseases as well as on the collective burden due to diabetes and CKD. The analyses in this report are innovative for Australia. Previously, the Global Burden of Disease (GBD) 2013 study reported high fasting plasma glucose levels and low glomerular filtration rate—biomedical indicators of diabetes and CKD—as risk factors for diseases (GBD 2013 Risk Factors Collaborators 2015). The New Zealand burden of disease study also undertook diseases-as-risks analysis for some conditions, including diabetes and CKD (MOH 2012). Results from this report provide estimates of the collective burden of diabetes and CKD, which can give a more comprehensive picture on the health loss attributable to these conditions. These results could be used to inform population health monitoring and may assist in the development of chronic disease policy. Another report in this series looks at a range of modifiable vascular risk factors for dementia, including vascular diseases that act as risk factors for dementia—diabetes, stroke, atrial fibrillation and CKD, and estimates their individual and combined contribution to the burden of dementia in Australia (AIHW 2016d). This first chapter provides background information on diabetes and CKD and their impact on other diseases. It then describes the ‘burden of disease’ approach used in the subsequent analysis and finally summarises the content of the following chapters.
1.2 Background
More about diabetes and CKD Definitions of diabetes and CKD for the purposes of analyses in this report are provided in Box 1.1. A range of behavioural and metabolic risk factors increase the risk of a person developing type 2 diabetes and CKD, such as tobacco smoking, overweight and obesity and high blood pressure. Similarly, diabetes and CKD increase the risk of developing other chronic diseases. There is evidence to suggest that diabetes can increase the risk of coronary heart disease, stroke, kidney disease, cancer and dementia (Cheng et al. 2014; Emerging Risk Factors Collaboration 2010; Prince et al. 2014 Vigneri et al. 2009). For example, diabetes is now the leading cause of treated end-stage kidney disease (ESKD) in Australia, accounting for 1 in 3
2 Diabetes and chronic kidney disease as risks for other diseases new cases in 2011 (ANZDATA 2013). ESKD is associated with large excesses of cardiovascular and all-cause mortality (Afkarian et al. 2013), and development of cardiovascular diseases such as coronary heart disease and heart failure (Tong & Stevenson 2007).
Box 1.1: Diabetes and chronic kidney disease Diabetes Diabetes is a chronic disease marked by high levels of glucose in the blood. It is caused either by the inability to produce insulin, or by the body not being able to use insulin effectively, or both. Diabetes manifests as a high level of sugar in the blood. The main types of diabetes are: • type 1 diabetes—an autoimmune disease that usually has onset in childhood or early adulthood • type 2 diabetes—largely preventable, usually associated with lifestyle factors and with later onset • gestational diabetes—when higher than normal blood glucose is diagnosed during pregnancy (AIHW 2014). In this report, gestational diabetes was not considered as a risk factor for other linked diseases; diabetes includes type 1 and type 2 but not gestational diabetes. This is consistent with the ABDS 2011, where gestational diabetes was not included under diabetes. Chronic kidney disease (CKD) CKD refers to all conditions of the kidney, lasting at least 3 months, where a person has had evidence of kidney damage and/or reduced kidney function. Evidence of kidney damage manifests as either urinary protein or albumin (a type of protein that is a more sensitive and specific marker of kidney disease), blood in the urine, or scarring detected by imaging tests. CKD is usually categorised into 5 stages (1 to 5, where stage 1 is least kidney damage and stage 5 is most kidney damage) according to the level of kidney function and evidence of kidney damage, indicated by biological markers such as blood or protein in the urine. In the ABDS 2011, CKD stages were defined using eGFR (estimated glomerular filtration rate) results for kidney function. In this study, CKD as a risk factor included stage 3 to 5, consistent with the ABDS 2011.
Diabetes and chronic kidney disease as risks for other diseases 3
Figure 1.2 illustrates possible causal pathways of the development of diabetes and CKD and associated diseases.
Risk factors for diabetes (type 2) Risk factors for chronic kidney disease ● Tobacco smoking ● Tobacco smoking ● Overweight and obesity ● Overweight and obesity ● Insufficient physical activity ● Insufficient physical activity ● High blood pressure ● High blood pressure ● Excess alcohol intake ● Inadequate fruit and vegetable intake ● Inadequate fruit and vegetable intake ● Impaired glucose regulation ● Impaired glucose regulation ● Diabetes
Diabetes Chronic kidney disease
Impacts of diabetes on the body ● Platelet hyperactivity Impacts of CKD on the body ● Endothelial dysfunction ● Platelet dysfunction ● Micro-angiopathy ● Abnormal metabolism ● Cardiac and renal fibrosis ● Cardiac and renal fibrosis ● Hypertension ● Hypertension ● Abnormal metabolism ● Anaemia ● Oxidative stress/inflammation ● Oxidative stress/inflammation
Diseases associated with diabetes ● Coronary heart disease Diseases associated with CKD ● Heart failure ● Coronary heart disease ● Stroke ● Heart failure ● Peripheral vascular disease ● Stroke ● Cancer ● Peripheral vascular disease ● Cognitive disorders, e.g. dementia ● Cognitive disorders, e.g. dementia ● Chronic kidney disease
Sources: AIHW 2014, Montemarano et al. 2013, Vulesevic et al. 2014. Figure 1.2: An example of possible causal pathways of diabetes and CKD to associated diseases
What is burden of disease? Burden of disease analyses assesses and compares the health impact of different diseases, conditions or injuries (referred to as ‘diseases’ for simplicity) and risk factors on a population. It captures the impact of both living with the disease and dying prematurely. The ABDS 2011 quantified the fatal and non-fatal effects of these diseases in a consistent manner so that they can then be combined into a summary measure of health called the DALY—disability-adjusted life years. The DALY combines the estimates of years of life lost due to premature death (YLL) and years lived in ill health or with disability (YLD) to count
4 Diabetes and chronic kidney disease as risks for other diseases the total years of healthy life lost from disease and injury. These and other key terms relating to burden of disease analyses are defined in Box 1.2. Taking all diseases into account, this health loss represents the difference between the current health status of the population and the ideal situation where everyone lives a long life, free of disease. Burden of disease estimates capture both the quantity and quality of life, and reflect the magnitude, severity and impact of disease and injury within a population in the given year. The analysis also estimates the contribution of various risk factors to health loss, known as the attributable burden. Attributable burden reflects the direct relationship between a risk factor (diabetes or CKD in this report) and a disease outcome (linked disease). It is the amount by which disease burden would be reduced if exposure to the risk factor (including diseases as risks) had been avoided or reduced to the lowest possible exposure. For detailed information about the most recent ABDS, and further information on the methods used to calculated disease burden, please refer to Australian Burden of Disease Study: impact and causes of illness and death in Australia 2011 (AIHW 2016a) and Australian Burden of Disease Study 2011: methods and supplementary material (AIHW 2016b).
1.3 Structure of this report This report provides a comprehensive analysis of the direct and indirect burden of diabetes and CKD using a ‘diseases-as-risks’ approach. The structure of this report is: • Chapter 2 describes the methods used in this report to estimate the direct and indirect burden of diabetes and CKD. • Chapter 3 summarises the results of analyses to estimate the impact of diabetes on the burden of linked diseases. • Chapter 4 summarises the results of analyses to estimate the impact of CKD on the burden of linked diseases. • Chapter 5 provides commentary on the implications of the findings, strengths and limitations of the study and concluding remarks. • Appendixes A and B provide information on the selection of effect sizes used in this report for the association between diabetes and linked diseases, and CKD and linked diseases, respectively. • Appendix C describes the methods used in scenario modelling where the direct and indirect burden due to diabetes and CKD were extrapolated to 2020 under 2 different scenarios. • Appendix D provides information on the data sources used for the diabetes and CKD estimates presented in this report.
Diabetes and chronic kidney disease as risks for other diseases 5
Box 1.2: Key terms used in this report attributable burden: The disease burden attributed with a particular risk factor. It is the reduction in fatal and non-fatal burden that would have occurred if exposure to the risk factor had been avoided (or, more precisely, had been at its theoretical minimum). collective burden: The sum of the direct and indirect burden. comparative risk assessment: The process for estimating the burden of disease attributable to selected risk factors. It involves 5 key steps: selection of risk–outcome pairs; estimation of exposure distribution; estimation of effect sizes; choice of theoretical minimum risk exposure level; and finally the calculation of attributable burden. DALY (disability-adjusted life years): Measure (in years) of healthy life lost, either through premature death defined as dying before the expected life span (YLL) or, equivalently, through living with ill health due to illness or injury (YLD). direct burden: In burden of disease analysis, it is the burden calculated to capture the main disabling consequences of the disease. For example, the direct diabetes burden includes diabetic nephropathy, neuropathy and retinopathy. diseases-as-risks: Diseases act as risk factors for other diseases. To fully account for the health loss attributable to diseases-as-risks requires that their ‘indirect’ burdens be calculated and then added to their ‘direct’ burdens in order to estimate their collective burdens. effect size: A statistical measure of the strength of the relationship between 2 variables (in this context, between a risk exposure and a disease outcome), expressed, for example, as a relative risk. hazard ratio (HR): Hazard ratio has similar interpretation to relative risk. It is the ratio of the probability of an event (e.g. disease outcome) in the exposed group to the probability in the control group. Hazard ratios differ from relative risks in that the latter are cumulative over an entire study, using a defined endpoint, while the former represent instantaneous risk at some particular time period during the study. indirect burden: In burden of disease analysis, where the disease of interest is considered to be a risk factor (that is, disease-as-risk) for linked diseases, it is the burden attributable to the disease-as-risk for linked diseases. For example, diabetes is considered to be a risk factor for coronary heart disease, stroke, dementia and other diseases, so the indirect burden is the burden attributable to diabetes for these linked diseases. linked disease: Many diseases can act as risk factors for developing certain diseases; for example, diabetes is associated with increased risk of developing coronary heart disease. The disease in association is a linked disease to the risk factor. meta-analysis: A statistical technique for combining findings from previous independent studies. It provides a quantitative estimate of the overall effect of an intervention or variable on a defined outcome, giving due weight to the size of the different studies included. population attributable fraction (PAF): The proportion (fraction) of a disease, illness, disability or death in a population that can be attributed to a particular risk factor or combination of risk factors. (continued)
6 Diabetes and chronic kidney disease as risks for other diseases
Box 1.2 (continued): Key terms used in this report odds ratio (OR): Odds ratio is a measure of association which compares the odds of disease in those exposed to the odds of disease in those unexposed. relative risk (RR): The risk of an event relative to exposure, calculated as the ratio of the probability of the event occurring in the exposed group to the probability of it occurring in the non-exposed group. A relative risk of 1 implies no difference in risk; a RR <1 implies the event is less likely to occur in the exposed group; and a RR >1 implies the event is more likely to occur in the exposed group. risk factor (for health): Any factor that causes or increases the likelihood of a health disorder or other unwanted condition or event. YLD (years lived with disability): A measure of the years of what could have been a healthy life but were instead spent in states of less than full health. YLD represent non-fatal burden. YLL (years of life lost): Years of life lost due to premature death defined as dying before the ideal life span. YLL represent fatal burden. (See Glossary for a full list of definitions used in the report.)
Diabetes and chronic kidney disease as risks for other diseases 7
2 Methods
The disease burden attributable to diabetes or CKD was estimated using the comparative risk assessment methodology (see Box 1.2 for definition). This is the standard approach for burden of disease risk factor analysis globally (GBD 2013 Risk Factors Collaborators 2015). In this study, diabetes and CKD are the risk factors of interest. To measure the contribution of diabetes or CKD, the calculations use high-quality information on evidence of a causal association between diabetes or CKD and other diseases. The amount of extra risk of developing or dying from that disease caused by exposure to diabetes or CKD (that is, effect size) and the number of people in the population exposed to diabetes or CKD are also included in the calculation. Exposure to the risk factor disease is measured as the disease prevalence in the population. In this study, the steps followed were: • select linked diseases and the effect size of risk factors on linked disease • estimate the risk factor (diabetes or CKD) exposure and define the theoretical minimum risk exposure distribution • calculate the population attributable fraction • quantify the indirect diabetes or CKD burden. These steps are further explored in this chapter and form the structure of this chapter.
2.1 Defining linked diseases Diseases linked to diabetes or CKD were chosen for analysis if they met the following criteria: • sufficient evidence for a causal association between exposure and linked disease outcome based on high quality epidemiological studies • sufficient data to estimate population exposure to the risk factor • a plausible biological mechanism linking risk factor and linked disease • modifiable or preventable. For diabetes-as-risk, most literature pertained to type 2 diabetes. In this study, calculation of diabetes burden included type 1 and type 2, consistent with the ABDS 2011 which did not calculate direct burden by diabetes type. Further, this diseases-as-risk study included: diagnosed diabetes, where diabetes prevalence was obtained from self-reported data or measured biomedical data; and undiagnosed diabetes, where diabetes has not been diagnosed by a doctor, but an individual has plasma glucose levels that satisfy established criteria for diabetes. Twelve linked diseases to diabetes-as-risk were included: 3 cardiovascular diseases (coronary heart disease, stroke and peripheral vascular disease), 7 types of cancer (liver, pancreatic, bowel, breast, uterine, kidney and bladder), chronic kidney disease and dementia. See Appendix A for a detailed description of the selection of linked diseases for diabetes. For CKD-as-risk, most literature defined CKD by eGFR of less than 60 mL/min/1.73 m2 or by the levels of urinary protein or albumin. This definition of a reduced eGFR is consistent
8 Diabetes and chronic kidney disease as risks for other diseases with stages 3 to 5 of CKD used in the ABDS 2011. Four linked diseases were included: coronary heart disease, stroke, peripheral vascular disease and dementia. See Appendix B for a detailed description of the selection of linked diseases for CKD.
Criteria for inclusion of linked diseases For a linked disease outcome to be included, there must be convincing or probable evidence of a causal association following exposure (to diabetes or CKD), preferably from a meta-analysis or cohort study. This aligns with the World Cancer Research Fund criteria of convincing or probable evidence for a causal association (see Box 2.1).
Box 2.1: World Cancer Research Fund criteria for level of evidence The World Cancer Research Fund used a criterion for grading evidence to support a judgement of a relationship with cancer. Its grading system breaks down data sources into ‘convincing’, ‘probable’, ‘possible’ and ‘insufficient’ evidence (WCRF & AICR 2007). Convincing evidence describes a causal relationship that is ‘robust enough to be highly unlikely to be modified in the foreseeable future as new evidence accumulates’. Probable evidence suggests a causal relationship is often described and is unlikely to change with increased knowledge.
Evidence for inclusion of linked disease was categorised as convincing or probable based on the robustness and volume of studies demonstrating a relationship. Convincing evidence included linked diseases with a well-known causal relationship, or where numerous high-quality studies applicable to Australia demonstrated a causal relationship after adjusting for confounders. A probable level of evidence included linked diseases where a causal relationship had been identified by high-quality studies, but supporting evidence was not as robust as those categorised as convincing. The main reason for classifying as probable evidence was that a meta-analysis had not been conducted, or only a few high-quality studies were available for selection. Pancreatic cancer was an exception, with a strong causal relationship to diabetes demonstrated by a number of studies; however, there is uncertainty on the direction of the relationship suggested by 1 study, depending on the time period of developing both conditions. The association between diabetes and pancreatic cancer was therefore classified as probable level of evidence. The levels of evidence are further explained in Appendix A for diabetes as a risk and Appendix B for CKD as a risk.
Diseases excluded A linked disease was included in this study if it met the criteria of inclusion as mentioned above and if it was a specific disease captured in the ABDS 2011. Heart failure as a linked disease for diabetes and CKD is supported by a substantial body of research. However, it was not considered a separate disease in the ABDS 2011; instead the effects of heart failure were included as a consequence of a number of underlying cardiovascular diseases (such as coronary heart disease and rheumatic heart disease). As such, its burden was not estimated separately in the ABDS 2011, and therefore, it was not possible to include a separate calculation of heart failure burden in this study.
Diabetes and chronic kidney disease as risks for other diseases 9
Some other conditions, such as sleep apnoea and thyroid disease, may also be associated with diabetes (Baronea & Menna-Barretob 2011; Kadiyala et al. 2010). The burden of these linked diseases were not able to be included in this study because they were not captured as separate diseases in the ABDS 2011. While there is some evidence of a possible association between CKD and some cancers, cancer was not included as a linked disease for CKD in this study. This was because the literature suggests these associations can be bi-directional dependent on therapies used to treat CKD, and there was not enough convincing evidence for specific cancers to enable its inclusion (Stengel 2010).
2.2 Selection of effect size Burden of disease studies use effect size to measure the association between risk factors and disease outcomes. For the linked diseases included in this study, except for diabetes and its association with CKD, effect sizes were selected from published meta-analyses or applicable prospective studies (see tables A1 and B1). In this study, meaures of effect include relative risk (RR), odds ratio (OR) and hazard ratio (HR). For the association between diabetes and CKD, direct evidence from the Australia and New Zealand Dialysis and Transplant Registry (ANZDATA) was used to ascertain the proportion of people who have CKD due to diabetes (the population attributable fraction). The registry captures data for virtually all cases of treated ESKD in Australia and provides evidence to attribute CKD to diabetes. The proportions were modelled for people aged 90 and over due to limited data on the registry for these ages. Each effect size was applied to both fatal and non-fatal burden. Depending on the method used in the primary analysis, effect sizes may have been adjusted for confounders as well as other diseases included in this study as described in appendix tables A2 and B2. The selection of effect sizes for diabetes or CKD and its linked disease is described in detail in Appendixes A and B.
2.3 Population exposure: prevalence of diabetes or CKD Definitions used to calculate exposure to the risk factors diabetes and CKD are described in Table 2.1. Further details on exposure data sources are in Appendix D. Exposure was aligned as best as possible with the evidence of causal association in the studies where effect sizes were sourced. For this study, exposure was treated as a dichotomous categorical variable: that is, the prevalence of those who have diabetes or CKD compared with the prevalence of those who do not have diabetes or CKD. The proportion of those who do not have diabetes or CKD is referred to as the theoretical minimum risk exposure level or the proportion ‘not exposed’. Those who have diabetes or CKD are the ‘exposed’ population and at risk of linked diseases.
10 Diabetes and chronic kidney disease as risks for other diseases
Table 2.1: Risk factor population exposure definition, 2011
Risk factor Definition of exposure
Prevalence of diagnosed and undiagnosed diabetes but with established Diabetes high plasma glucose level (type 1 and type 2).
Chronic kidney disease Prevalence of persons with CKD, stages 3 to 5.
Australian population distributions of diabetes and CKD by age and sex were sourced directly from the ABDS 2011 (AIHW 2016b). Diabetes and CKD definitions (Table 2.1) were used to guide data extraction. The ABDS 2011 derived the proportions of people who have and do not have diabetes or CKD using the finest possible increments from the original data source. To reduce the impact of survey error, the data were extracted at a level where the relative standard error was 25% or less. For further information on the methods used to derive disease prevalence, please refer to the Australian Burden of Disease Study 2011: methods and supplementary material (AIHW 2016b). The prevalence of a risk factor influences the attributable burden. That is, the attributable burden may be greater for a risk factor with a small effect size and high prevalence compared with a risk factor with a large effect size and low prevalence. Therefore, this analysis is dependent on accurate prevalence estimates for diabetes or CKD.
2.4 Calculation of population attributable fractions Population attributable fractions (PAFs) determine the proportion of a particular disease that could have potentially been avoided if the population had never been exposed to a risk factor (Box 2.2). The calculation of PAFs requires the input of: • the effect size, that is the RR, OR or HR, of the risk factor on the outcome of interest, and • the prevalence of exposure in the population (P). The PAF is calculated as:
( 1) = ( 1) + 1 푃 푒푓푓푒푐푡 푠푖푧푒 − 푃퐴퐹 Attributable burden (AB) is calculated 푃as:푒푓푓푒푐푡 푠푖푧푒 − = ×
Where, C = the direct퐴퐵 burden푃퐴퐹 (DALY)퐶 of a specific outcome, such as stroke.
Diabetes and chronic kidney disease as risks for other diseases 11
Box 2.2: Example of how a population attributable fraction is applied to the population In the population, a proportion of all liver cancer is due to diabetes. This proportion is estimated using a population attributable fraction (PAF) which takes into account the number of people in each age group, say for males aged 50–54. It uses the number of males in this age range who have diabetes (for example 9%) and the size of the association between the risk factor and the linked disease (liver cancer). In this case, the relative risk is 2.31 (selected from published meta-analyses). This is calculated using the following formula: ( 1) = ( 1) + 1 푃 푅푅 − 푃퐴퐹 Using this formula for diabetes and liver cancer푃 푅푅, −we get: 0.09(2.31 1) = 0.09(2.31 1) + 1 − 푃퐴퐹 = 0.105− In 2011, 2,840 DALY were estimated for 푃퐴퐹liver cancer in Australian men aged 50–54. Attributable burden is an estimate of the amount of this liver cancer burden that is due to diabetes. This is calculated by multiplying the population attributable fraction and the linked disease burden. AB = PAF x DALY AB = 0.105 x 2,840 AB = 298 Therefore, 298 DALY from liver cancer in males aged 50–54 were attributable to diabetes. Note that this is an example and the calculations are done separately for each age group and sex.
2.5 Disease-as-risks analysis: calculating indirect burden In burden of disease analysis, each disease is defined to be mutually exclusive to every other disease in the study. This enables the burden from all diseases to be summed to estimate the total burden in the population. As a result, the burden calculated in the ABDS 2011 only captured the direct burden of each disease. However, some diseases act as risk factors for other diseases and to fully account for the health loss attributable to diseases-as-risks requires that their ‘indirect’ or additional burden be calculated. Their direct and indirect burden can then be added to estimate their collective burden. For example, in the ABDS 2011, microvascular complications of diabetes (such as diabetic nephropathy, neuropathy and retinopathy) are considered a main disabling consequence of diabetes and captured in the direct diabetes burden, but the macrovascular complications of diabetes (such as coronary heart disease, peripheral vascular disease and stroke) are captured under the more direct cardiovascular condition, not included under diabetes. Diabetes is considered to be a risk factor for coronary heart disease, stroke, dementia and other diseases, so the burden from these linked diseases due to diabetes (the ‘indirect’
12 Diabetes and chronic kidney disease as risks for other diseases diabetes burden) can be added to the direct diabetes burden to estimate the collective diabetes burden without any overlap. When interpreting the results, it is important to note that collective burden of diabetes and CKD were calculated independently and it is not possible to add them together due to overlaps between these risk factors.
2.6 Scenario modelling Scenario modelling was used to explore the impact of changes in the prevalence of diabetes and CKD on the burden of linked diseases, as well as the collective burden for these 2 diseases under 2 scenarios: if the prevalence rate of diabetes or CKD continues its increasing trend, or if it remains steady to 2020. These scenarios provide an indication of the amount of burden that may be avoided if the current rise of diabetes or CKD prevalence is halted, compared with the amount of burden if the current trends continue. The year 2020 was chosen because it aligns with the National Strategic Framework for Chronic Conditions and the World Health Organization’s Global Action Plan for the Prevention and Control of Non-communicable Diseases 2013–2020 (WHO 2013). Results from both methods were compared with 2011 burden to estimate the difference in the impact between the 2 scenarios. Detailed information on the methods used for scenario modelling is in Appendix C.
Diabetes and chronic kidney disease as risks for other diseases 13
3 Diabetes estimates
This chapter describes the diabetes burden estimates as reported in the ABDS 2011 (referred to as the ‘direct burden’) and the burden from linked diseases due to diabetes (referred to as the ‘indirect burden’). The list of diseases included is outlined in Table A1.
3.1 Diabetes burden in the ABDS 2011 In 2011, diabetes accounted for 2.3% of the total burden of diseases and injuries in Australia (AIHW 2016a). Overall, diabetes was the 12th leading cause of burden, equating to 101,653 DALY. A higher proportion of diabetes burden was evident in males (58%) compared with females (42%) (Table 3.1). The diabetes burden had a similar proportion of fatal and non-fatal burden (53% fatal; 47% non-fatal).
Table 3.1: Diabetes direct burden, by sex, 2011
DALY YLL YLD
Number % Number % Number %
Males 59,298 58.3 31,114 57.5 28,183 59.3
Females 42,356 41.7 22,996 42.5 19,360 40.7
Persons 101,653 100.0 54,110 100.0 47,543 100.0
Source: AIHW burden of disease database, 2011.
In both males and females, diabetes burden varied by age (Figure 3.1). Males experienced a greater amount of burden in most age groups. In males, burden increased steadily to age 69, reaching a peak of 8,461 DALY, and then decreased rapidly after age 84. Females experienced more burden than males from age 85 onwards, probably influenced by the relative longevity in females. In females, the number of DALY increased with age—but not as steeply as seen in males—peaking at age 80–84 (5,169 DALY).
DALY (number) 9,000 Males 8,000 Females 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 1–4 5–9 100+ 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85–89 90–94 95–99 10–14 Under 1 Under Age group (years)
Source: AIHW burden of disease database, 2011. Figure 3.1: Diabetes direct burden, by age and sex, 2011
14 Diabetes and chronic kidney disease as risks for other diseases
3.2 What is the collective burden of diabetes? The diabetes burden reported in the ABDS 2011 is referred to as the direct burden. The burden of other diseases attributable to diabetes is referred to as the indirect burden. The sum of the direct and indirect burden is the collective diabetes burden. Table 3.2 shows the number of DALY for diabetes due to direct and indirect burden, and its total collective burden. There were 101,653 DALY estimated due to the direct impact of diabetes in 2011. Additional 88,332 DALY were estimated due to the indirect burden of diabetes on linked diseases. Collectively, diabetes was responsible for 189,985 DALY. The collective burden due to diabetes was almost twice (1.9 times) the direct diabetes burden reported in the ABDS 2011. By including indirect diabetes burden, the burden of diabetes for males increased by 84% and for females it increased by 90%.
Table 3.2: Collective burden (DALY) of diabetes, by direct and indirect burden and sex, 2011
Direct Indirect Collective
Males 59,298 50,045 109,343 Females 42,356 38,288 80,644
Persons 101,653 88,332 189,985
Note: Numbers may not sum to the total due to rounding.
Source: AIHW analysis of burden of disease database, 2011. The indirect and direct diabetes burden differed by age (Figure 3.2). The indirect burden occurred at a later age than direct burden, because the linked diseases are generally conditions that progress over an extended period and predominantly affect older people. In both males and females, the indirect burden was responsible for over 50% of the collective diabetes burden from age 75 onwards.
(a) Males (b) Females
DALY (number) DALY (number) 20,000 20,000
Direc t Direc t 15,000 Indirect 15,000 Indirect
10,000 10,000
5,000 5,000
0 0 0 95+ 0 95+ Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 3.2: Collective burden (DALY) of diabetes, by direct and indirect burden, by age and sex, 2011
Diabetes and chronic kidney disease as risks for other diseases 15
DALY (%) DALY (%) 100 100
80 80
60 60
40 40
20 20
0 0 0 95+ 0 95+ Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 3.2 (continued): Collective burden (DALY) of diabetes, by direct and indirect burden, by age and sex, 2011
3.3 How much burden from linked diseases is due to diabetes? The indirect diabetes burden was estimated as the number of DALY due to linked diseases. Overall, the indirect diabetes burden was responsible for 88,332 DALY (Table 3.3). The burden attributable to diabetes was highest in coronary heart disease (38,852 DALY), followed by stroke (18,730 DALY) and CKD (8,945 DALY). These 3 diseases were responsible for 75% of the indirect diabetes burden, with almost half (44%) of the total indirect diabetes burden from coronary heart disease. The indirect diabetes burden varied by sex, with males experiencing 31% more burden than females (50,045 DALY compared with 38,288 DALY). This is due to both overall diabetes prevalence and the burden of some linked diseases such as coronary heart disease, being higher among males.
16 Diabetes and chronic kidney disease as risks for other diseases
Table 3.3: Attributable burden due to diabetes, by linked disease and sex, 2011
Males Females Persons
Proportion of Proportion of Proportion of Attributable Attributable Attributable Linked disease linked linked linked DALY (a) DALY (a) DALY (a) disease (%) disease (%) disease (%)
Cardiovascular diseases
Coronary heart 23,539 10.4 15,313 12.7 38,852 11.2 disease
Stroke 8,864 13.5 9,866 13.9 18,730 13.7
Peripheral vascular 509 10.9 262 7.0 771 9.2 disease
Chronic kidney 5,164 24.0 3,781 17.9 8,945 21.0 disease
Dementia(b) 3,597 6.5 4,421 4.6 8,018 5.3
Cancer
Liver cancer 2,926 13.5 729 9.6 3,655 12.4
Pancreatic cancer 2,397 9.7 1,258 6.4 3,655 8.2
Bowel cancer 2,078 3.9 915 2.3 2,993 3.2
Breast cancer — — 875 1.4 875 1.4
Kidney cancer 570 4.6 170 3.1 739 4.2
Uterine cancer — — 610 8.0 610 8.0
Bladder cancer 401 3.4 88 2.1 489 3.1
Total 50,045 38,288 88,332
(a) The ‘Proportion of linked disease (%)’ column is the attributable DALY due to diabetes divided by the total direct linked disease burden estimated in the ABDS of that row. Numbers may not sum to the total due to rounding.
(b) The burden of dementia due to diabetes was estimated only in people aged 65 and over, because the association is in late life dementia only.
Source: AIHW analysis of burden of disease database, 2011. The proportion of disease burden that is attributable to diabetes was estimated for each linked disease (Table 3.3 and Figure 3.3). This is derived from the number of DALY attributable to diabetes divided by the direct burden of each linked disease. For example, we estimated 38,852 DALY of coronary heart disease can be attributable to diabetes; this is 11% of the direct burden of coronary heart disease (346,651 DALY). In males, 24% of the chronic kidney disease burden, 13% of the liver cancer burden and 13% of the stroke burden was due to diabetes. Diabetes was also responsible for 11% of the peripheral vascular disease burden and 10% of the coronary heart disease burden. In females, diabetes was responsible for 18% of the chronic kidney disease burden, 14% of the stroke burden and 13% of coronary heart disease burden (Table 3.3, Figure 3.3). Comparing male and female proportions of attributable DALY by disease, males had a larger proportion of underlying disease burden attributable to diabetes, with the exception of coronary heart disease and stroke. The largest absolute difference was seen in CKD, liver cancer, peripheral vascular disease and pancreatic cancer.
Diabetes and chronic kidney disease as risks for other diseases 17
Uterine and breast cancer were estimated for the female population only. Diabetes was attributable to 8.0% and 1.4% of the uterine and breast cancer burden, respectively.
Linked disease
Chronic kidney disease Stroke Coronary heart disease Liver cancer Peripheral vascular disease Pancreatic cancer Dementia Kidney cancer Bow el cancer Bladder cancer Uterine cancer Males Breast cancer Females
0 5 10 15 20 25 Proportion of direct DALY (%)
Source: AIHW analysis of burden of disease database, 2011. Figure 3.3: Proportion of burden attributable to diabetes, by linked disease and sex, 2011
Was the indirect burden due to fatal or non-fatal outcomes? In the ABDS 2011, the direct burden of diseases associated with diabetes was predominantly fatal, with the exception of dementia burden, where it was mainly non-fatal. This influenced the proportion of fatal and non-fatal outcomes for the indirect diabetes burden. The indirect diabetes burden was mostly due to fatal burden, with the exception of dementia attributable to diabetes (54% fatal; 46% non-fatal). The proportion of indirect diabetes burden that was fatal differed by linked disease (Figure 3.4). The proportion of fatal burden for linked cancers ranged between 89% of the breast cancer burden attributable to diabetes to 99% of the liver cancer burden attributable to diabetes. The proportion of fatal burden for linked cardiovascular diseases ranged from 79% of the coronary heart disease burden attributable to diabetes to 88% of the stroke burden attributable to diabetes. Similarly, 71% of the burden of CKD attributable to diabetes was due to fatal outcomes.
18 Diabetes and chronic kidney disease as risks for other diseases
Linked disease
Liver cancer Pancreatic cancer Kidney cancer Bladder cancer Bow el cancer Uterine cancer Breast cancer Stroke PV D CHD CKD Dementia
0 25 50 75 100 Proportion of indirect DALY (%) Fatal Non-fatal
Note: PVD = Peripheral vascular disease; CHD = Coronary heart disease; CKD = Chronic kidney disease.
Source: AIHW analysis of burden of disease database, 2011. Figure 3.4: Proportion of indirect diabetes burden, by linked disease due to fatal and non-fatal outcomes, 2011
The proportion of indirect diabetes burden that was fatal differed by age (Figure 3.5). The proportion of fatal outcomes was higher for younger age groups—responsible for 81–85% of the indirect diabetes burden under age 55. This decreased to 79% between ages 65 and 74 and increased slightly to 81% for age 85 and over.
Age group (years) Under 45 45–54 55–64 65–74 75–84 85–94 95+
0 25 50 75 100 Proportion of indirect DALY (%) Fatal Non-fatal
Source: AIHW analysis of burden of disease database, 2011. Figure 3.5: Proportion of indirect diabetes burden due to fatal and non-fatal outcomes, by age, 2011
3.4 Indirect diabetes burden by linked disease This section looks at the impact of diabetes on individual linked diseases by age and sex. The indirect burden was measured by the number and proportion of DALY from the disease that was caused by the risk factor—in this case diabetes. Where possible, similar scales have
Diabetes and chronic kidney disease as risks for other diseases 19 been used for graphs across linked diseases to depict the relative size of the burden estimates reported and to aid interpretation.
Cardiovascular disease estimates
Coronary heart disease Diabetes contributed to 11% of the coronary heart disease burden in 2011, equating to 38,852 DALY (Table 3.3). Males experienced a greater number of DALY (23,539), compared with females (15,313). Males experienced most of the coronary heart disease DALY due to diabetes between ages 65 and 84 (Figure 3.6a). Females experienced most of the attributable burden at a later age, with the majority occurring between ages 75 and 94. Overall, the proportion of coronary heart disease burden due to diabetes increased with age (Figure 3.6b). The proportion of attributable burden peaked at 14% in males at ages 65–74 and decreased to 8.7% at age 95 and over. In females the attributable burden gradually increased with age, and was higher than males from ages 75 onwards, peaking at 18% at age 95 and over.
(a) (b)
DALY (number) DALY proportion (%) 8,000 25 Males 20 6,000 Females 15 4,000 10 2,000 5
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under 45 Under Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 3.6: Burden (DALY) of coronary heart disease attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011
Stroke In total, 14% of the stroke burden was attributable to diabetes or 18,730 DALY (Table 3.3). Males experienced a slightly lower number of DALY (8,864), compared with females (9,866), but a similar proportion of the burden was evident in both sexes (13.5% males; 13.9% females). The stroke burden due to diabetes increased steeply with increasing age in males, peaking at 3,411 DALY between ages 75 and 84 (Figure 3.7a). From age 85, the number of attributable DALY decreased dramatically. Females also experienced an increase in DALY with increasing age, but a lower number of DALY compared with males before age 75. Unlike
20 Diabetes and chronic kidney disease as risks for other diseases males, the number of DALY in females remained high and was similar between ages 75–84 and 85–94. In males and females from age 55 onwards, between 10 and 20% of the stoke burden was attributable to diabetes (Figure 3.7b). In males, the proportion was highest at ages 65–74, where 18% of stroke burden was due to diabetes. In females, this proportion was highest at age 95 and over (20%).
(a) (b)
DALY (number) DALY proportion (%) 8,000 25 Males 20 6,000 Females 15 4,000 10 2,000 5
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under Under 45 Under
Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 3.7: Burden (DALY) of stroke attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011
Peripheral vascular disease Diabetes contributed to 9.2% of the peripheral vascular disease burden in 2011 (771 DALY; Table 3.3). Males experienced two-thirds of the attributable burden (509 DALY), with 11% of the peripheral vascular disease burden in males due to diabetes. This proportion was lower in females (7.0%). Similar to coronary heart disease, the attributable burden increased with age and was higher in males in all ages, compared with females, except at age 95 and over (Figure 3.8a). The proportion of attributable burden peaked at 14% in males at ages 65–74 and decreased to 8.7% at age 95 and over (Figure 3.8b). In females, the attributable burden gradually increased with age, but was much lower in all ages, with the exception of those aged 95 and over.
Diabetes and chronic kidney disease as risks for other diseases 21
(a) (b)
DALY (number) DALY proportion (%) 1,000 25 Males 800 20 Females 600 15
400 10
200 5
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under Under 45 Under Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 3.8: Burden (DALY) of peripheral vascular disease attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011
Cancer estimates
Liver cancer In total, 12% of the liver cancer burden was attributable to diabetes (3,655 DALY; Table 3.3), most of which (2,926 DALY; 80%) was experienced by males. This is influenced by the higher prevalence of both diabetes and liver cancer in males compared with females. The proportion of liver cancer burden attributable to diabetes was higher in males (13%) compared with females (9.6%). Males experienced most of the liver cancer attributable burden between ages 55 and 74, peaking at 972 DALY between ages 65 and 74 (Figure 3.9a). Females experienced attributable burden at a later age, with the majority occurring between ages 65 and 84. Between 10 and 20% of the liver cancer burden after age 65 was attributable to diabetes (Figure 3.9b). In males, the proportion was highest at ages 65–74, where 20% of liver cancer burden was due to diabetes. In females, this proportion was highest at age 95 and over (16%).
22 Diabetes and chronic kidney disease as risks for other diseases
(a) (b)
DALY (number) DALY proportion (%) 1,000 25 Males 800 20 Females 600 15
400 10
200 5
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under Under 45 Under Age group (years) Age group (years)
Note: For age group 95+, liver cancer attributable to diabetes was <10 DALY.
Source: AIHW analysis of burden of disease database, 2011. Figure 3.9: Burden (DALY) of liver cancer attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011
Pancreatic cancer In 2011, 8.2% of the pancreatic cancer burden was attributable to diabetes, equating to 3,655 DALY (Table 3.3). Males experienced almost two-thirds (66%) of the pancreatic cancer burden due to diabetes. The proportion of pancreatic cancer burden attributable to diabetes was also higher in males (9.7%), compared with females (6.4%). In males, the number of attributable DALY increased steeply with age, peaking at 942 DALY between ages 65 and 74 (Figure 3.10a). From age 75 onwards, this decreased dramatically. In females, the burden increased gradually and was experienced at a later age compared with males. The majority of the burden in females occurred between ages 65 and 84. The proportion of pancreatic cancer burden due to diabetes also increased with age and peaked at 13% in males aged 65–74 and 11% in females aged 95 and over (Figure 3.10b).
(a) (b)
DALY (number) DALY proportion (%) 1,000 25 Males 800 20 Females 600 15
400 10
200 5
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under Under 45 Under Age group (years) Age group (years) Note: For age group 95+, pancreatic cancer attributable to diabetes was <10 DALY. Source: AIHW analysis of burden of disease database, 2011. Figure 3.10: Burden (DALY) of pancreatic cancer attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011
Diabetes and chronic kidney disease as risks for other diseases 23
Bowel cancer In 2011, 3.2% of the bowel cancer burden was attributable to diabetes, equating to 2,993 DALY (Table 3.3). Males experienced 69% of the bowel cancer attributable burden due to diabetes. The proportion of bowel cancer burden attributable to diabetes was less than 5% in both males (3.9%) and females (2.3%). In males, bowel cancer burden due to diabetes increased steeply with age, peaking at 849 DALY between ages 65 and 74 (Figure 3.11a). In females, the number of attributable DALY increased slightly with age, peaking at ages 75–84. The proportion of attributable burden peaked at 5.3% in males aged 65–74 and 4.3% in females aged 95 and over (Figure 3.11b).
(a) (b)
DALY (number) DALY proportion (%) 1,000 25 Males 800 20 Females 600 15
400 10
200 5
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under 45 Under Age group (years) Age group (years)
Note: For age group 95+, bowel cancer attributable to diabetes was <20 DALY.
Source: AIHW analysis of burden of disease database, 2011. Figure 3.11: Burden (DALY) of bowel cancer attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011
Uterine and breast cancer in females In 2011, 8.0% of the uterine cancer burden and 1.4% of the breast cancer burden in females was attributable to diabetes, equating to 610 and 875 DALY, respectively (Table 3.3). The burden of breast cancer due to diabetes was estimated only in women aged 45 and over, because studies found its association with diabetes was mainly in post-menopausal women (Boyle et al. 2012; Larsson et al. 2007). In females, the number of attributable DALY for uterine and breast cancer mostly occurred between ages 55 and 84 (Figure 3.12a). Between 9 and 11% of uterine cancer over age 65 was attributable to diabetes (Figure 3.12b). The proportion of attributable burden of breast cancer due to diabetes peaked at 2.6% in women aged 95 and over (Figure 3.12b).
24 Diabetes and chronic kidney disease as risks for other diseases
(a) (b)
DALY (number) DALY proportion (%) 1,000 25 Breast Breast 800 Uterine 20 Uterine 600 15
400 10
200 5
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under Under 45 Under Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 3.12: Burden (DALY) of breast and uterine cancers attributable to diabetes, by (a) number and (b) proportion of burden, by age and cancer type, 2011
Kidney and bladder cancer Diabetes contributed to 4.2% of kidney and 3.1% of the bladder cancer burden in 2011 (Table 3.3). This equated to 739 and 489 DALY, respectively. Males experienced 77% of the attributable diabetes burden from kidney cancer and 82% from bladder cancer. This compares with 23% and 18%, respectively, for females. This is influenced by higher direct burden of both kidney and bladder cancer in males compared with females. Attributable burden increased with age for kidney cancer, peaking between ages 65 and 74 (258 DALY). The attributable burden in bladder cancer increased at a later age, with both linked cancers showing a similar number of DALY between ages 75 and 84 (Figure 3.13a). The proportion of attributable burden peaked at ages 65–74 for kidney cancer (5.8%) and bladder cancer (3.7%) (Figure 3.13b).
Diabetes and chronic kidney disease as risks for other diseases 25
(a) (b)
DALY (number) DALY proportion (%) 1,000 25 Kidney Kidney 800 20 Bladder Bladder 600 15
400 10
200 5
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under Under 45 Under Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 3.13: Burden (DALY) of kidney and bladder cancer attributable to diabetes, by (a) number and (b) proportion of burden, by age and cancer type, 2011
Dementia estimates Diabetes contributed to 5.3% of the dementia burden at age 65 and over, equating to 8,018 DALY (Table 3.3). The burden of dementia due to diabetes was estimated only in people aged 65 and over, because the association is in late life dementia only. Males experienced a slightly lower number of DALY (3,597) compared with females (4,421), but the proportion of dementia burden attributable to diabetes was higher in males (6.5% compared with 4.6% in females). Males experienced most of the attributable burden between ages 75 and 84, peaking at 1,719 DALY (Figure 3.14a). Females experienced attributable burden at a later age, peaking between ages 85 and 94 (1,964 DALY). The proportion of dementia burden attributable to diabetes decreased with age in males, from 8.5% at ages 65–74 to 5.1% at age 95 and over (Figure 3.14b). In females, this increased with age, from 4.7% at ages 65–74 to 6.5% at age 95 and over.
26 Diabetes and chronic kidney disease as risks for other diseases
(a) (b)
DALY (number) DALY proportion (%) 2,500 25 Males 2,000 Females 20
1,500 15
1,000 10
500 5
0 0 65–74 75–84 85–94 95+ 65–74 75–84 85–94 95+ Age group (years) Age group (years)
Note: Attributable burden was not calculated for ages under 65, because the association is with dementia in late-life (age 65 and over).
Source: AIHW analysis of burden of disease database, 2011. Figure 3.14: Burden (DALY) of dementia attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011
Chronic kidney disease estimates Diabetes contributed to 21% of the CKD burden, equating to 8,945 DALY (Table 3.3). Males experienced 57% of the attributable burden (5,164 DALY), and the proportion of CKD burden attributable to diabetes was higher in males (24% compared with 18% in females). Males and females experienced most of the attributable burden between ages 75 and 84 (1,492 DALY and 947 DALY, respectively) (Figure 3.15a), and 21–30% of CKD between ages 45–74 was due to diabetes (Figure 3.15b).
(a) (b)
DALY (number) DALY proportion (%) 2,500 35 Males 30 2,000 Females 25 1,500 20 1,000 15 10 500 5 0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under Under 45 Under Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 3.15: Burden (DALY) of chronic kidney disease attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011
Diabetes and chronic kidney disease as risks for other diseases 27
3.5 Estimated diabetes burden in 2020 under 2 different scenarios Table 3.4 presents the indirect and direct diabetes burden in 2011, compared with the diabetes burden in 2020 under 2 different scenarios. These scenarios provide an indication of the amount of burden that may be avoided if the current rise of diabetes is halted, compared with the amount of burden if the current trend continues (see Appendix D for detailed methods used). For the direct burden, Scenario A describes the impact of the annual rate of change in fatal and non-fatal burden rates between 2003 and 2011 continued to 2020. Scenario B describes the impact of population growth and ageing, where the diabetes fatal and non-fatal burden rates remain stable to 2020. For the indirect burden, Scenario A describes the impact of the annual rate of change in diabetes prevalence between 2003 and 2011 continued to 2020. Scenario B describes the impact of the diabetes prevalence rate remaining stable to 2020. The 2011 burden rates for each linked disease were assumed to remain the same to 2020 for ease of calculation. The annual rate of change in prevalence and fatal burden between 2003 and 2011 was derived for each sex and age group (see tables C1 and C2). The annual rate of change in non-fatal burden between 2003 and 2011 was calculated at the sequela level for each sex and age group. Sequelae are consequences associated with the disease, that is undiagnosed and diagnosed diabetes (see Appendix D). The results of the scenario analyses are presented below. If the rate of change observed between 2003 and 2011 in diabetes prevalence, fatal and non-fatal burden continued to 2020: • The estimated indirect diabetes burden is projected to be 180,000 DALY in 2020, which is twice as high as the indirect burden in 2011 (Table 3.4). • The estimated direct diabetes burden is projected to be 130,000 DALY in 2020, fatal burden is expected to be 6% higher, and non-fatal burden 52% higher, than the direct burden in 2011 (Table 3.4). • The estimated collective diabetes burden is projected to be 310,000 DALY in 2020, which 1.6 times as high as the collective burden in 2011 (Table 3.4). If the current rise in diabetes is halted (that is, the prevalence, fatal and non-fatal burden rates remain stable from 2011 to 2020), then by 2020: • the estimated collective diabetes burden is projected to be 244,000 DALY, which is 1.3 times as high as the collective burden in 2011 (and reflects population increase and ageing alone) (Table 3.4). In comparing the 2 scenarios, if the current rise in diabetes is halted compared with if the current trend continues, then by 2020: • 21% of future collective diabetes burden could be avoided (Table 3.4) • 36% of future indirect diabetes burden could be avoided, with reductions ranging from 32% for CKD to 40% for bladder cancer and dementia (Table C5).
28 Diabetes and chronic kidney disease as risks for other diseases
Table 3.4: Estimated diabetes burden in 2020 under different exposure scenarios Scenario B compared with Scenario A Scenario B Scenario A
Continued trend in diabetes prevalence Stable diabetes prevalence rate to 2020 DALY that would be avoided in 2020 rate to 2020
Percentage change Percentage change (b) DALY in 2011 DALY in 2020 (a) DALY in 2020 (a) DALY number DALY (%) from 2011 DALY from 2011 DALY
Indirect burden total 88,332 180,084 103.9 115,788 31.1 64,297 35.7
Direct diabetes burden
Fatal (YLL) 54,110 57,382 6.0 69,315 28.1 –11,933 –20.8
Non-fatal (YLD) 47,543 72,431 52.3 59,233 24.6 13,197 18.2
Total (DALY)(c) 101,653 129,812 27.7 128,548 26.5 1,264 1.0
Collective diabetes (d) 189,985 309,896 63.1 244,337 28.6 65,561 21.2 burden
(a) Percentage change is the change in burden between 2020 and 2011 divided by the burden in 2011 for each scenario.
(b) Percentage difference is the difference in attributable DALY in 2020 between Scenario A and B divided by the attributable DALY in 2020 for Scenario A.
(c) Numbers may not sum to the total due to rounding.
(d) Collective diabetes burden is the sum of indirect diabetes burden and the total (DALY) for each scenario.
Source: AIHW analysis of burden of disease database, 2011.
Diabetes and chronic kidney disease as risks for other diseases 29
3.6 Scenario differences by age
Indirect diabetes burden The indirect diabetes burden in 2011 occurred mainly in older ages, because the linked diseases progress over an extended period of time and become more prevalent with age. Figure 3.16 shows the indirect diabetes burden by age in 2020 under the 2 different scenarios. In both scenarios, the indirect burden estimated for 2020 was highest in people aged 75–84. Under both scenarios, the indirect DALY due to diabetes was relatively similar in those aged under 45 and 95 and over. If the current rise in diabetes is halted, 34% of the future indirect burden due to diabetes in people aged 45–64, and 36% in those aged 65–94, could be avoided in 2020.
Estimated indirect DALY in 2020 60,000
50,000
40,000
30,000
20,000
10,000
0 Under 45 45–54 55–64 65–74 75–84 85–94 95+ Age group (years)
Indirect DALY from stable diabetes prevalence rate Indirect DALY from continued diabetes prevalence trend
Note: ‘Stable diabetes’ is the scenario where the rise of diabetes prevalence is halted; ‘continued diabetes’ is the scenario where the current prevalence trends continue to rise. The stable diabetes prevalence rate is based on 2011 prevalence rate.
Source: AIHW analysis of burden of disease database, 2011. Figure 3.16: Estimated indirect diabetes burden (DALY) in 2020 under different exposure scenarios, by age
30 Diabetes and chronic kidney disease as risks for other diseases
4 Chronic kidney disease estimates
This chapter examines the CKD estimates as reported in the ABDS 2011 (direct burden) and the burden in linked diseases due to CKD (indirect burden). The linked diseases include coronary heart disease, stroke, dementia and peripheral vascular disease.
4.1 Chronic kidney disease burden in the ABDS 2011 In 2011, CKD accounted for 0.9% of the total burden of diseases and injuries in Australia, equating to 42,574 DALY (AIHW 2016a). The number of DALY was similar by sex (Table 4.1). Nearly three-quarters (72%) of CKD burden was due to fatal outcomes.
Table 4.1: Chronic kidney disease direct burden, by sex, 2011
DALY YLL YLD
Number % Number % Number %
Males 21,490 50.5 15,680 51.2 5,810 48.7
Females 21,084 49.5 14,965 48.8 6,119 51.3
Persons 42,574 100.0 30,645 100.0 11,929 100.0
Note: Numbers may not sum to the total due to rounding.
Source: AIHW burden of disease database, 2011.
CKD burden increased with age, with the exception of burden in infants (Figure 4.1). Males experienced a greater amount of burden, before age 85. In males, burden increased steadily to age 84, reaching a peak of 3,575 DALY. Females experienced more burden than males from age 85 onwards, peaking at 80–84 (3,300 DALY).
DALY (number) 4,000 Males 3,500 Females 3,000
2,500
2,000
1,500
1,000
500
0 1–4 5–9 100+ 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85–89 90–94 95–99 Under 1 Under Age group (years) Source: AIHW burden of disease database, 2011. Figure 4.1: Chronic kidney disease direct burden, by age and sex, 2011
Diabetes and chronic kidney disease as risks for other diseases 31
4.2 What is the collective burden of chronic kidney disease? Table 4.2 shows the number of DALY for CKD due to direct and indirect burden, and its total collective burden. There were 42,574 DALY estimated due to the direct impact of CKD in 2011. Additional 46,866 DALY were estimated due to the indirect burden of CKD on linked diseases. Collectively, CKD was responsible for 89,460 DALY. The collective burden due to CKD was twice (2.1 times) the direct CKD burden reported in the ABDS 2011. By including the indirect CKD burden, the burden of CKD for males doubled, and for females it more than doubled (ratio of 2.2).
Table 4.2: Collective burden (DALY) of chronic kidney disease, by direct and indirect burden and sex, 2011
Direct Indirect Collective
Males 21,490 21,152 42,642 Females 21,084 25,732 46,816
Persons 42,574 46,886 89,460
Note: Numbers may not sum to the total due to rounding.
Source: AIHW analysis of burden of disease database, 2011. The age disparity between indirect and direct burden in both sexes is shown in Figure 4.2. Males experienced most of the collective burden between ages 75 and 89. The proportion of indirect burden increased with age, with 10% of the collective burden between ages 50 and 59 being indirect burden, compared with 60–65% after age 75, reflecting the onset of cardiovascular diseases and dementia in late life. Females experienced most of the collective burden in later life due to linked diseases— primarily because females live longer than males. Indirect CKD burden was responsible for less than 10% of the collective CKD burden up to age 65. From age 80 onwards, this proportion increased to 68–73% of the collective burden.
(a) Males (b) Females
DALY (number) DALY (number) 10,000 10,000
Direc t Direc t 7,500 Indirect 7,500 Indirect
5,000 5,000
2,500 2,500
0 0 0 95+ 0 95+ Age group (years) Age group (years) Source: AIHW analysis of burden of disease database, 2011. Figure 4.2: Collective burden (DALY) of chronic kidney disease, by direct and indirect burden, by age and sex, 2011
32 Diabetes and chronic kidney disease as risks for other diseases
DALY (%) DALY (%) 100 100
80 80
60 60
40 40
20 20
0 0 0 95+ 0 95+ Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 4.2 (continued): Collective burden (DALY) of chronic kidney disease, by direct and indirect burden, by age and sex, 2011
4.3 How much burden from other diseases is due to chronic kidney disease? The indirect disease burden due to CKD was responsible for 46,886 DALY (Table 4.3). This is slightly more than the direct burden estimated as part of the ABDS 2011 (42,574 DALY). The attributable burden was highest in coronary heart disease (22,728 DALY)—responsible for almost half (48%) of the total attributable burden—followed by dementia (12,678 DALY; 27%) and stroke (9,859 DALY; 21%). The overall CKD attributable burden varied by sex, with males experiencing a slightly lower number of DALY, compared with females. This was mainly due to females experiencing a greater amount of dementia burden.
Table 4.3: Attributable burden due to chronic kidney disease, by linked disease and sex, 2011
Males Females Persons Proportion of Proportion of Proportion of Attributable Attributable Attributable Linked disease linked linked linked DALY (a) DALY (a) DALY (a) disease (%) disease (%) disease (%)
Coronary heart 12,186 5.4 10,541 8.7 22,728 6.6 disease
Dementia(b) 4,067 7.3 8,611 9.0 12,678 8.4
Stroke 4,063 6.2 5,796 8.2 9,859 7.2
Peripheral 836 17.9 784 21.0 1,621 19.2 vascular disease
Total 21,152 25,732 46,886
(a) The ‘Proportion of linked disease (%)’ column is the attributable DALY due to CKD divided by the total direct linked disease burden estimated in the ABDS of that row. Numbers may not sum to the total due to rounding.
(b) The burden of dementia due to CKD was estimated only in people aged 65 and over, because the association is in late life dementia only.
Source: AIHW analysis of burden of disease database, 2011.
Diabetes and chronic kidney disease as risks for other diseases 33
The proportion of disease burden that is attributable to CKD was estimated for each linked disease (Table 4.3 and Figure 4.3). This is derived from the number of DALY attributable to CKD divided by the direct burden of each linked disease. For example, we estimated 22,728 DALY of coronary heart disease can be attributable to CKD; this is 6.6% of the direct burden of coronary heart disease (346,651 DALY). CKD contributed the greatest proportion of attributable burden for peripheral vascular disease: 18% in males and 21% in females. For the remaining linked diseases, CKD attributed 5–7% of the disease burden in males and 8–9% in females. A larger proportion of disease burden was attributable to CKD in females in every linked disease (Table 4.3, Figure 4.3). This is mainly due to an increased amount of burden occurring in older age groups in females, compared with males.
Linked disease
Peripheral vascular disease
Dementia
Stroke
Males Coronary heart disease Females
0 5 10 15 20 25 Proportion of direct DALY (%)
Source: AIHW analysis of burden of disease database, 2011. Figure 4.3: Proportion of burden attributable to chronic kidney disease, by linked disease and sex, 2011
In the ABDS 2011, the direct burden of diseases associated with CKD was predominantly fatal, with the exception of dementia burden, where it was mainly non-fatal. This influenced the proportion of fatal and non-fatal outcomes for the indirect CKD burden. The proportion of indirect CKD burden was mostly fatal; however, this varied by linked disease (Figure 4.4). Over 80% of indirect burden from linked cardiovascular disease was due to fatal burden: stroke (89%), peripheral vascular disease (85%) and coronary heart disease (80%) (Figure 4.4). A similar proportion of fatal and non-fatal burden was evident in the dementia burden attributable to CKD (54% fatal; 46% non-fatal).
34 Diabetes and chronic kidney disease as risks for other diseases
Linked disease
Stroke
PV D
CHD
Dementia
0 25 50 75 100 Proportion of indirect DALY (%) Fatal Non-fatal Note: PVD = Peripheral vascular disease; CHD = Coronary heart disease.
Source: AIHW analysis of burden of disease database, 2011. Figure 4.4: Proportion of indirect chronic kidney disease burden, by linked disease due to fatal and non-fatal outcomes, 2011
The proportion of indirect CKD burden due to fatal and non-fatal outcomes differed by age (Figure 4.5). Fatal outcomes were responsible for 85–88% of the indirect CKD burden under age 55. After this age, fatal outcomes were responsible for around three-quarters of the indirect CKD burden.
Age group (years) Under 45 45–54 55–64 65–74 75–84 85–94 95+
0 25 50 75 100 Proportion of indirect DALY (%) Fatal Non-fatal
Source: AIHW analysis of burden of disease database, 2011. Figure 4.5: Proportion of indirect chronic kidney disease burden due to fatal and non-fatal outcomes, by age, 2011
4.4 Indirect CKD burden by linked disease The impact of CKD on linked diseases is discussed individually for each disease: coronary heart disease, stroke, dementia and peripheral vascular disease. The indirect burden was measured by the number and proportion of DALY from the disease that was caused by the risk factor—in this case CKD. Where possible, similar scales have been used for graphs across linked diseases to depict the relative size of the burden estimates reported and to aid interpretation.
Diabetes and chronic kidney disease as risks for other diseases 35
Coronary heart disease CKD contributed to 6.6% of the coronary heart disease burden in 2011, equating to 22,728 DALY (Table 4.3). Males experienced a slightly greater number of DALY (12,186), compared with females (10,541 DALY). Males experienced nearly half of the total attributable burden between ages 75 and 84, peaking at 5,181 DALY (Figure 4.6a). Females experienced most of the attributable burden at a later age, peaking at 4,866 DALY between ages 85 and 94. From age 75 onwards, CKD was responsible for 10–14% of coronary heart disease burden (Figure 4.6b). The proportion of attributable burden increased with age, and peaked at 12% in males aged 95 and over, and 14% in females aged 95 and over.
(a) (b)
DALY (number) DALY proportion (%) 6,000 40 Males Females 30 4,000 20 2,000 10
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under 45 Under Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 4.6: Burden (DALY) of coronary heart disease attributable to chronic kidney disease, by (a) number and (b) proportion of burden, by age and sex, 2011
Stroke CKD contributed to 7.2% of the stroke burden in 2011 (9,859 DALY; Table 4.3). Females experienced 59% of the total attributable burden (5,796 DALY). Nearly all stroke burden due to CKD occurred after age 65 (Figure 4.7a). Males experienced the majority of the attributable burden between ages 75 and 94. The attributable burden peaked later for females: it was highest in older age groups, peaking at 2,734 DALY between ages 85 and 94, and is considerably higher than males. From age 75 onwards, CKD was responsible for 10–13% of stroke burden (Figure 4.7b). The proportion of attributable burden increased with age, and peaked at 11% in males aged 95 and over, and 13% in females aged 95 and over.
36 Diabetes and chronic kidney disease as risks for other diseases
(a) (b)
DALY (number) DALY proportion (%) 6,000 40 Males Females 30 4,000 20 2,000 10
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under Under 45 Under Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 4.7: Burden (DALY) of stroke attributable to chronic kidney disease, by (a) number and (b) proportion of burden, by age and sex, 2011
Dementia CKD contributed to 8.4% of the dementia burden age 65 and over, equating to 12,678 DALY (Table 4.3). The burden of dementia due to CKD was estimated only in people aged 65 and over, because the association is with late life dementia only. Females experienced 68% of the total attributable burden. Nearly all dementia burden due to CKD occurred after age 65, coinciding with dementia increasing with age (Figure 4.8a). Males experienced the majority of their attributable burden between ages 75 and 94. Females experienced most of their attributable burden at a later age, peaking at 4,585 DALY between ages 85 and 94. CKD contributed to 9–11% of dementia burden after age 75 (Figure 4.8b). The proportion of attributable burden increased with age, and peaked at 10% in males and 11% in females aged 95 and over.
Diabetes and chronic kidney disease as risks for other diseases 37
(a) (b)
DALY (number) DALY proportion (%) 6,000 40 Males Females 30 4,000
20
2,000 10
0 0 65–74 75–84 85–94 95+ 65–74 75–84 85–94 95+ Age group (years) Age group (years)
Note: Attributable burden was not calculated in ages under 65, because the association with dementia is with late-life (aged 65 and over).
Source: AIHW analysis of burden of disease database, 2011. Figure 4.8: Burden (DALY) of dementia attributable to chronic kidney disease, by (a) number and (b) proportion of burden, by age and sex, 2011
Peripheral vascular disease CKD contributed to 19% of the peripheral vascular disease burden in 2011 (1,621 DALY; Table 4.3). Males experienced 52% of the attributable burden (836 DALY). In males, the burden occurred at an earlier age than for females, peaking at 330 DALY at ages 75–84 (Figure 4.9a). In females, the attributable burden peaked between ages 85 and 94 (367 DALY). From age 75 onwards, CKD contributed to 27–33% of the burden from peripheral vascular disease (Figure 4.9b).
(a) (b)
DALY (number) DALY proportion (%) 1,000 40 Males 800 Females 30 600 20 400 10 200
0 0 95+ 95+ 45–54 55–64 65–74 75–84 85–94 45–54 55–64 65–74 75–84 85–94 Under 45 Under Under 45 Under Age group (years) Age group (years)
Source: AIHW analysis of burden of disease database, 2011. Figure 4.9: Burden (DALY) of peripheral vascular disease attributable to chronic kidney disease, by (a) number and (b) proportion of burden, by age and sex, 2011
38 Diabetes and chronic kidney disease as risks for other diseases
4.5 Estimated CKD burden in 2020 under 2 different scenarios Table 4.4 presents the indirect and direct CKD burden in 2011, compared with the CKD burden in 2020 under 2 different scenarios. These scenarios provide an indication of the amount of burden that may be avoided if the current rise of CKD is halted, compared with the amount of burden if the current trend continues (see Appendix D for the detailed methods used). For the direct burden, Scenario A describes the impact of the annual rate of change in fatal and non-fatal burden rates between 2003 and 2011 continued to 2020. Scenario B describes the impact of population growth and ageing, where the CKD fatal and non-fatal burden rates remain stable to 2020. For the indirect burden, Scenario A describes the impact of the annual rate of change in CKD prevalence between 2003 and 2011 continued to 2020. Scenario B describes the impact of the CKD prevalence rate remaining stable to 2020. The 2011 burden rates for each linked disease were assumed to remain the same to 2020 for ease of calculation. The annual rate of change in prevalence and fatal burden between 2003 and 2011 was derived for each sex and age group (see tables C3 and C4). The annual rate of change in non-fatal burden between 2003 and 2011 was calculated at the sequela level for each sex and age group. Sequelae are consequences associated with the disease, such as stages 3 to 5 of CKD (see Appendix D). The results of the scenario analyses are presented below. If the rate of change observed between 2003 and 2011 in CKD prevalence, fatal and non-fatal burden continued to 2020: • The estimated indirect CKD burden is projected to be 67,400 DALY in 2020, which is 1.4 times as high as the indirect burden in 2011 (Table 4.4). • The estimated direct CKD burden is projected to be 55,800 DALY in 2020; fatal burden is expected to be 25% higher, and non-fatal burden 45% higher, than the direct burden in 2011 (Table 4.4). • The estimated collective CKD burden is projected to be 123,000 DALY in 2020, which is 1.4 times as high as the collective burden in 2011 (Table 4.4). If the current rise in CKD is halted (that is, the prevalence, fatal and non-fatal burden rates remain stable from 2011 to 2020), then by 2020: • the estimated collective CKD burden is projected to be 117,000 DALY, which is 1.3 times as high as the collective burden in 2011 (and reflects population increase and ageing alone). In comparing the 2 scenarios, if the current rise in CKD is halted compared with if the current trend continues, then by 2020: • 4.8% of future CKD collective burden could be avoided (Table 4.4) • 7.5% of future indirect CKD burden could be avoided, with reductions ranging between 6.0% for dementia and peripheral vascular disease, to 8.2% for coronary heart disease (Table C6).
Diabetes and chronic kidney disease as risks for other diseases 39
Table 4.4: Estimated chronic kidney disease burden in 2020 under different exposure scenarios Scenario B compared with Scenario A Scenario B Scenario A
Continued trend in CKD prevalence rate Stable CKD prevalence rate to 2020 DALY that would be avoided in 2020 to 2020
Percentage change Percentage change (b) DALY in 2011 DALY in 2020 (a) DALY in 2020 (a) DALY number DALY (%) from 2011 DALY from 2011 DALY
Indirect CKD burden 46,886 67,410 43.8 62,342 33.0 5,068 7.5
Direct CKD burden
Fatal (YLL) 30,645 38,421 25.4 39,780 29.8 -1,359 -3.5
Non-fatal (YLD) 11,929 17,338 45.3 15,178 27.2 2,160 12.5
Total (DALY)(c) 42,574 55,758 31.0 54,958 29.1 800 1.4
Collective CKD (d) 89,440 123,168 37.7 117,300 31.1 5,868 4.8 burden
(a) Percentage change is the change in burden between 2020 and 2011 divided by the burden in 2011 for each scenario.
(b) Percentage difference is the difference in attributable DALY in 2020 between Scenario A and B divided by the attributable DALY in 2020 for Scenario A.
(c) Numbers may not sum to the total due to rounding.
(d) Collective CKD burden is the sum of indirect CKD burden and the total (DALY) for each scenario.
Source: AIHW analysis of burden of disease database, 2011.
40 Diabetes and chronic kidney disease as risks for other diseases 4.6 Scenario differences by age
Indirect CKD burden Figure 4.10 shows the indirect CKD burden in 2020 by age under the 2 different scenarios. In both scenarios, the indirect burden estimated for 2020 was highest in ages 75–94. Under both scenarios, the indirect DALY due to CKD was similar in all age groups except for ages 75–84. If the increase in the rate of CKD prevalence is halted, 15% of the future indirect burden due to CKD in people aged 65–84 could be avoided in 2020.
Estimated indirect DALY in 2020 30,000
25,000
20,000
15,000
10,000
5,000
0 Under 45 45–54 55–64 65–74 75–84 85–94 95+ Age group (years)
Indirect DALY from stable CKD prevalence rate Indirect DALY from continued CKD prevalence trend
Note: ‘Stable CKD’ is the scenario where the rise of CKD prevalence is halted; ‘continued CKD’ is the scenario where the current prevalence trends continue to rise. The stable CKD prevalence rate is based on 2011 prevalence rate.
Source: AIHW analysis of burden of disease database, 2011. Figure 4.10: Estimated indirect chronic kidney disease burden (DALY) in 2020 under different exposure scenarios, by age
Diabetes and chronic kidney disease as risks for other diseases 41
5 Discussion
5.1 Key findings This study found that the burden due to diabetes and CKD is around twice as high when taking into account the indirect burden of their linked diseases. By including the indirect burden due to these diseases with their direct burden, the total burden of disease due to diabetes increased by 84% for males and by 90% for females. For CKD, this increased by 98% for males and by 122% for females. Scenario analyses suggested that around one-fifth (21%) of future diabetes burden and 5% of future CKD burden could be avoided if the prevalence rate of these 2 diseases is maintained at 2011 levels to 2020. This proportion is higher for diabetes than CKD because: firstly, diabetes was responsible for more direct burden than CKD in 2011 (101,653 DALY compared with 42,574 DALY); and secondly, because the rate of change used to model current and future trends in disease prevalence and burden was greater for diabetes than CKD (see Appendix C). In terms of indirect burden, the greatest gains are expected to be made in those aged 65–94 for diabetes, and 65–84 for CKD, where around 36% of diabetes attributable burden and 15% of CKD attributable burden could be avoided by 2020 if the prevalence rate of these diseases is maintained at 2011 levels.
5.2 Contribution to evidence base and policy analysis This study provides insight into the impact of diabetes and CKD on a number of other chronic diseases by looking at them as risk factors for these other diseases. It does this by quantifying the collective burden—the sum of the direct and indirect impact or burden associated with health loss from these conditions. The impact of diabetes and CKD on population health is emphasised by their collective burden. When the indirect burden of diabetes and CKD was added to their direct burden, the collective burden is 1.9 and 2.1 times, respectively, higher than their direct burden. This highlights the contribution of diabetes and CKD to disease burden, both in their own right and as risk factors for other diseases, such as cardiovascular diseases and dementia. There is evidence that reducing the prevalence of diabetes or CKD could, at a population level, reduce the prevalence of a broader range of chronic diseases. For example, several studies have shown the observed prevalence of age-specific dementia is decreasing, possibly due to improvements in cardiovascular health, including hypertension and diabetes over past decades (Norton et al. 2014). There is consistent evidence that optimal glycaemic control (along with other modifiable risk factors) could reduce cardiovascular risk in patients with type 2 diabetes (Martín-Timón et al. 2014). Scenario analysis in this study also showed that if the current rise of diabetes or CKD prevalence is halted, chronic disease burden may be reduced across a range of diseases. Given the age of exposure, duration and association with linked diseases, strategies aimed at preventing or delaying the onset of diabetes and CKD across the life-course have the potential to achieve major health gains. Early detection of risk factors and disease has the
42 Diabetes and chronic kidney disease as risks for other diseases potential to reduce or delay complications and consequently other chronic conditions, which would result in significant health benefits for the population. Results from this study provide substantial details about the association between diabetes and CKD and their linked diseases. The information about the extent of these relationships across sex, age groups and for different diseases may be used to better target disease prevention and management policies.
5.3 New Zealand comparison Disease-as-risk analysis was previously undertaken as part of the New Zealand Burden of Disease Study (MOH 2012). The New Zealand study used the same comparative risk assessment methodology as used in the Australian Burden of Disease study and this report, but assumptions for the inclusion of linked diseases and effect sizes differed. The New Zealand study had a more conservative approach to selecting linked diseases and effect sizes. Effect sizes were only included if they were greater than 1.4, and the linked disease was responsible for greater than 1% of the total burden in at least 1 population group. These criteria were considered to exclude some linked diseases that could be measured based on a convincing or probable level of evidence, and therefore were not applied in this study. As a result, a greater number of linked diseases were included in this Australian analysis. Despite these differences, the proportion of attributable burden by linked diseases common to both studies can be compared. Coronary heart disease and stroke were identified in both studies as linked diseases for diabetes. Table 5.1 shows the comparison between results in this study and indirect burden estimates from the New Zealand study. Overall, a similar proportion of attributable burden due to diabetes was identified for coronary heart disease and stroke combined, with differences in the contribution of each.
Table 5.1: Comparison of direct and indirect diabetes burden between Australia and New Zealand diseases-as-risks analyses
ABDS estimates New Zealand estimates
Direct burden Indirect burden Direct burden Indirect burden
Proportion Proportion Attributable Attributable Linked disease of linked of linked DALY (a) DALY (a) disease(%) disease(%)
Coronary heart 346,651 38,852 11.2 89,159 12,900 14.5 disease
Stroke 136,771 18,730 13.7 37,688 3,600 9.6
Total 483,422 57,582 11.9 126,847 16,500 13.0
(a) The ‘Proportion of linked disease (%)’ column is the attributable DALY due to diabetes divided by the total direct linked disease burden estimated in the ABDS of that row. Numbers may not sum to the total due to rounding.
Source: AIHW analysis of burden of disease database, 2011, MOH 2012. Differences were also evident between studies for the indirect burden of CKD. The New Zealand study reported that one-third of the collective CKD burden was due to indirect burden. Our Australian study found the indirect CKD burden contributes approximately one-half of the collective CKD burden. This could be due to a greater number of linked diseases identified in this study contributing to the collective CKD burden.
Diabetes and chronic kidney disease as risks for other diseases 43
5.4 Limitations
Literature review and linked diseases Data are not available to obtain a complete understanding of the causal pathways and associated probabilities of developing linked diseases due to diabetes and CKD in Australia. The estimates of attributable burden used in this study rely on the best available effect size estimates from recent high-quality studies and meta-analyses. The studies included here as sources of effect sizes were deemed to be relevant for Australia and demonstrated a sufficient association between diabetes or CKD and their linked disease (based on a probable or convincing level of evidence). The exception was for CKD and its association with dementia, which was assessed as a possible/probable level of evidence. Further research is likely to update the effect sizes used for this association and may identify other possible linked diseases for diabetes and CKD. Meta-analysis was typically chosen over other study designs for effect size estimates to provide greater statistical power and an ability to extrapolate to the affected general Australian population. However, a limitation inherent in meta-analysis of published studies is the possibility of publication bias. Where meta-analyses were based on prospective studies, or case-control studies (which are susceptible to recall and selection bias), there may be issues with confounding factors that may explain the observed associations reported. However, many studies used in this study adjusted for potential confounders and the association was similar for studies that controlled for body mass index, physical inactivity and alcohol consumption, and for studies that did not adjust for these variables. There are several limitations that should be considered when interpreting results for the burden of diabetes. First, most of the literature used to determine effect sizes pertained to type 2 diabetes, with only limited literature available about type 1 diabetes as a risk factor for other diseases. The effect size was applied to exposure data which included both type 1 and type 2 diabetes. Because type 1 diabetes (which accounts for 5–10% of all diagnosed cases of diabetes) may not be a risk factor for linked diseases, the magnitude of the relationship between diabetes and linked diseases may have been somewhat overestimated. There were also limited studies in which the analysis provided results for type 2 diabetes by insulin status. We therefore could not examine whether attributable burden may differ for those using insulin to manage their type 2 diabetes. In selecting the literature for effect sizes of linked diseases to diabetes or CKD, where possible, the definitions of disease outcome in studies selected were similar to the definitions used in estimating the disease burden in the ABDS 2011 (tables A2 and B2). This was the case for most linked diseases except for peripheral vascular disease for which the definition used in the studies selected for effect sizes differ to that used in the ABDS 2011. Its effect size was based on an objective measure—Ankle Brachial Index, while the ABDS 2011 defined its disease consequence as intermittent claudication due to peripheral vascular disease. It was not possible to ascertain the relationship between the objective measure and self-reported symptoms. The indirect burden estimates for peripheral vascular disease reported in this study should therefore be interpreted with caution. Further, the effect size selected for the association between diabetes and stroke from the Emerging Risk Factors Collaboration (2010) referred to ischaemic stroke. Our study inferred this effect size to all stroke subtypes, as the majority of burden estimates in the ABDS 2011 was due to ischaemic stroke. This may have overestimated the indirect burden of stroke due to diabetes.
44 Diabetes and chronic kidney disease as risks for other diseases
The proportion of CKD that was attributable to diabetes used in this study was based on data from the ANZDATA. Using direct evidence from an Australian registry was considered to be more appropriate than using effect size from the literature for this linked disease pair. Because the ANZDATA is specific to persons with treated ESKD (stage 5 CKD), one assumption and limitation in using this data was that it was assumed that the proportion of people with stage 5 CKD due to diabetes was the same as for those with stage 3 or stage 4 CKD. Further details are provided in Appendix D.
Common risk factors It is important to consider common risk factors, such as smoking, obesity, physical inactivity and poor nutrition, are potential confounders to the associations of diabetes or CKD and their linked diseases. For example, for tobacco smoking, being a risk factor for diabetes and cardiovascular diseases, some people may develop both conditions independently due to their individual risk factor profile, rather than developing cardiovascular diseases due to having diabetes (AIHW 2012b). Scenario analysis was undertaken to estimate the extent of burden that could be reduced if the prevalence rates of diabetes and CKD remain stable compared with an increasing trend to 2020 (Appendix C). This was based on modelling the flow-on effects of controlling the direct burden due to diabetes and CKD, which then have an impact on the indirect burden of the linked diseases. However, it is acknowledged that, in treating or preventing chronic diseases, it is often important to target associated risk factors to limit the diseases’ development or progression (AIHW 2012b). Strategies that target common risk factors—such as smoking, obesity, physical inactivity and poor nutrition—are another way that the burden of diabetes and CKD might be reduced.
Estimating trends in diabetes and CKD prevalence for scenario modelling In order to estimate projected burden due to diabetes and CKD in 2020 used in the scenario analyses, information on current trends in the prevalence of these diseases was required. This was sourced from the ABDS 2011, which included estimates of prevalence for 2003 and 2011. Although biomedical data from the 2011–12 Australian Health Survey (AHS) was used to estimate prevalence of diagnosed diabetes and stage 3 and 4 CKD in 2011, comparable information was not available from earlier health surveys because they did not include biomedical components. As such, the 2003 estimates for diabetes were modelled using trends in self-report data from previous health surveys, which assumes that changes in self-reported diabetes follows a similar rate of change to diagnosed diabetes. For stage 5 CKD, ANZDATA was used to inform trends in prevalence. For stage 3 and 4 CKD, 2003 estimates were modelled using a ratio of the prevalence of treated ESKD sourced from ANZDATA to the prevalence of stage 3 CKD and stage 4 CKD sourced from the 2011–12 AHS. This assumes that this ratio was similar in 2003 as in 2011. These assumptions should be kept in mind when interpreting the results of the scenario modelling presented in this report. As mentioned earlier in the report, the 2011 burden rates for each linked disease were assumed to remain the same to 2020 for ease of calculation. This does not take into account potential improvements in disease outcomes in the future.
Diabetes and chronic kidney disease as risks for other diseases 45
5.5 Future directions This study presents a more detailed picture of the impact of diabetes and CKD in Australia. The results show that reducing diabetes or CKD could also reduce broader chronic disease burden, resulting in less health loss due to these 2 conditions and their linked diseases. Diabetes and CKD are also common chronic diseases among Aboriginal and Torres Strait Islander people. At present, published literature on the effect sizes specific to the Indigenous population is limited but effect sizes from this study could be applied as a proxy or re-selected from current literature. Differences in the indirect burden of diabetes and CKD between the Indigenous and total Australian populations would be mostly caused by differences in diabetes and CKD prevalence, and in the burden of linked diseases. However, other factors may also play a role, such as potential genetic predisposition—for example, smaller coronary arteries in women, which could lead to greater cardiovascular disease complications (Hiteshi et al. 2014)—and environmental factors. An analysis of the indirect diabetes and CKD burden in the Indigenous population would be an important area of work to progress in future burden of disease studies. As part of the ABDS 2011, the AIHW has developed a system that will allow estimates of burden of disease in Australia to be updated and kept current with emerging information. This offers potential to monitor and update the estimates included in this study as new evidence emerges about the association between diseases-as-risks and linked diseases, and as exposure to risk factors in the population changes over time.
5.6 Conclusion The impact of diabetes and CKD on population health is emphasised by their respective collective burden. When the indirect burden of diabetes and CKD was added to their direct burden, the collective burden was 1.9 and 2.1 times, respectively, higher than their direct burden. This highlights the contribution of diabetes and CKD to disease burden, both in their own right and as risk factors for other diseases, such as cardiovascular diseases. Results from this study provide information about the association between diabetes and CKD and their linked diseases. The collective burden of diabetes and CKD gives a more comprehensive picture on the health loss attributable to these conditions. The results in this study could be used to inform population health monitoring and may assist in the development of chronic disease policy for prevention initiatives.
46 Diabetes and chronic kidney disease as risks for other diseases
Appendix A: Selection of effect sizes for diabetes
In this report, diabetes includes type 1 and type 2, but not gestational, diabetes. This is consistent with the ABDS 2011. Selected literature on diabetes described in this appendix mainly refers to type 2 diabetes unless otherwise stated. Burden of disease studies use effect sizes to measure the strength of association between a risk factor (disease-as-risks in this study) and a linked disease. In this appendix, effect size measures the risk of developing the specific linked disease among those with and without diabetes. Effect sizes used in this study were identified following review of relevant literature and were restricted to studies with a prospective longitudinal design where the outcome was a clinically diagnosed linked disease. Sex-specific effect sizes were applied where possible. Selected studies showed that, if an individual has diabetes, there is an increased risk of developing a number of diseases: these are identified in Table A1. Appendix Table A1 also lists the effect sizes used in this study to estimate the indirect burden of diabetes and their sources. For further information on the selection of linked diseases and effect sizes see Chapter 2. A summary of selected studies is presented in Table A2.
Table A1: Effect size and sources for diabetes and linked disease pair analysis.
Linked disease Level of evidence Effect size (95% CI) Source of effect size
Coronary heart disease Convincing Males: HR 1.89 (1.73–2.06) Emerging Risk Factor Females: HR 2.59 (2.29–2.93) Collaboration 2010 Stroke (all types) Convincing Males: HR 2.16 (1.84–2.52) Emerging Risk Factor Females: HR 2.83 (2.35–3.40) Collaboration 2010 Peripheral vascular disease Convincing OR 1.88 (1.66–2.14) Fowkes et al. 2013 Dementia (all types; age 65+) Convincing RR 1.50 (1.33–1.70) Prince et al. 2014 Chronic kidney disease(a) — — ANZDATA 2011 Liver cancer Convincing RR 2.31 (1.87–2.84) Wang et al. 2012 Pancreatic cancer Probable OR 1.82 (1.66–1.89) Huxley et al. 2005 Bowel cancer Probable RR 1.30 (1.20–1.40) Larsson et al. 2005 Uterine cancer Probable RR 2.10 (1.75–2.53) Friberg et al. 2007 Breast cancer (age 45+) Convincing RR 1.20 (1.12–1.28) Larsson et al. 2007 Kidney cancer Convincing RR 1.40 (1.16–1.69) Bao et al. 2013 Bladder cancer Convincing RR 1.24 (1.08–1.42) Larsson et al. 2006
(a) Population attributable fraction for CKD was derived directly from the ANZDATA, 2011.
Diabetes and cardiovascular disease Diabetes is an established risk factor for coronary heart disease and stroke (AIHW 2014). Studies have shown that patients with diabetes have increased risk of stroke. For example, Liao et al. (2015) found the incidents of stroke in cohorts with and without diabetes were 10.1 and 4.5 per 1,000 persons-years, respectively. The increased risk of ischaemic stroke has been
Diabetes and chronic kidney disease as risks for other diseases 47
linked to pathophysiological changes seen in cerebral vessels of individuals with diabetes (Tuttolomondo et al. 2015). People with diabetes are at higher risk of developing atherosclerosis: the most common cause of peripheral vascular disease. Individuals with peripheral vascular disease have a much higher risk of heart attack or stroke. The pathophysiology of vascular disease in diabetes involves abnormalities in endothelial, vascular smooth muscle cells and platelet functions. In people with diabetes, the risk of peripheral vascular disease is increased by age, duration of diabetes and presence of peripheral neuropathy (Fowkes et al. 2013), as well as other risk factors such as smoking, hypertension and high cholesterol.
Diabetes and coronary heart disease There have been several cohort studies conducted supporting diabetes as an independent risk factor for coronary heart disease in people with diabetes (Selvin et al. 2005; Soedamah- Muthu et al. 2006; Tonelli et al. 2012). The effect size used in this study was drawn from a meta-analysis by the Emerging Risk Factors Collaboration (2010) where diabetes and the risk of vascular disease was assessed. It was a collaborative meta-analysis of 102 prospective studies, involving 1.27 million people. This was considered the most comprehensive meta-analysis, including the largest number of studies with an outcome of coronary heart disease and stroke by age and sex. The adjusted hazard ratio for the association between diabetes and coronary heart disease was 1.89 (95% CI 1.73–2.06) for males and 2.59 (95% CI 2.29–2.93) for females (Emerging Risk Factors Collaboration 2010). This was comparable to the relative risks reported in a meta-analysis by Peters et al. (2014a) of 2.16 for males and 2.82 for females. Both meta- analyses suggest that the effect of diabetes on the relative risk of coronary heart disease is greater in females than males.
Diabetes and stroke In this study, stroke as a linked disease referred to all types of stroke, not specific subtypes of stroke (that is, ischaemic and haemorrhagic stroke). Several cohort studies found diabetes is associated with increased risk of total and subtypes of stroke (Janghorbani et al. 2007; Liao et al. 2015; Najarian et al. 2006). A meta-analysis by Shou et al. (2015) further showed that stroke patients with diabetes had significantly higher stroke recurrence risks than those without diabetes. The effect size used in this study was also drawn from the Emerging Risk Factors Collaboration (2010). The adjusted hazard ratio for the association between diabetes and ischaemic stroke was 2.16 (95% CI 1.84–2.52) for males and 2.83 (95% CI 2.35–3.40) for females. This meta-analysis was selected as it included the largest number of studies with an outcome of stroke subtypes. The effect size was comparable to a meta-analysis by Peters et al. (2014b), where the relative risk of stroke associated with diabetes was 1.83 for males and 2.28 for females. The sex difference was consistent across total and subtypes of stroke.
Diabetes and peripheral vascular disease There is convincing evidence on diabetes as a risk factor for peripheral vascular disease. A meta-analysis by Fowkes et al. (2013) assessed the odds ratios for 15 risk factors for peripheral vascular disease for high-income and low-income countries, and included diabetes as a risk factor. The odds ratio (1.88, 95% CI 1.66–2.14) used for the association
48 Diabetes and chronic kidney disease as risks for other diseases between diabetes and peripheral vascular disease was drawn from this meta-analysis, using the effect size found for diabetes for high-income countries.
Diabetes and dementia There is strong and consistent evidence of a causal association between diabetes and dementia (Prince et al. 2014). Numerous prospective studies have demonstrated this as summarised in several systematic reviews (Biessels et al. 2006; Cukierman et al. 2005; Lu et al. 2009; Luchsinger 2010; Profenno et al. 2010). The association is also supported by recent meta-analyses (Cheng et al. 2012; Gudala et al. 2013; Lu et al. 2009; Ninomiya 2014; Prince et al. 2014). The relative risk used in this study is from Prince et al. (2014) because this was assessed to be the most up-to-date meta-analysis, and included the largest number of studies with an outcome of any dementia. In total, 11 studies were included in Prince et al. (2014) with a pooled relative risk of 1.50 for the association between any dementia and diabetes in late life (aged 65 and over). This relative risk was consistent when analysis was restricted to 7 studies that included both diagnosed and undiagnosed diabetes. The results from Prince et al. (2014) are in accordance with previous reviews and are comparable to other recent meta-analyses (Barnes & Yaffe 2011; Lu et al. 2009; Ninomiya 2014; Rönnemaa et al. 2011). There are several biologically plausible mechanisms through which diabetes could increase dementia risk. However, there is currently no clear consensus on the direct causal relationship (Ninomiya 2014; Prince et al. 2014). Diabetes is associated with atherosclerosis and stroke (Mankovsky & Ziegler 2004) which, in turn, increase the risk of dementia (Savva & Stephan 2010). In addition, defective binding of insulin to receptors in the brain may contribute to accumulation of amyloid, which is toxic to brain cells (Bedse et al. 2015). Brain imaging studies demonstrate an association between brain metabolism, consistent with Alzheimer’s disease and insulin resistance (Baker et al. 2011) and diabetes (Roberts et al. 2014). Finally, common genetic pathways may underlie the development of both diabetes and dementia (Akomolafe et al. 2006).
Diabetes and chronic kidney disease Direct evidence from the Australia and New Zealand Dialysis and Transplant Registry (ANZDATA) was used to ascertain a proportion of people in the database who have CKD due to diabetes (the population attributable fraction). The registry captures data for most cases of treated ESKD in Australia and provides evidence to attribute CKD to diabetes. The proportions were modelled for people aged 75 and over due to limited data on the registry for these ages. It is understood that people aged 85 and over are less likely to undergo renal replacement therapy, so the proportion due to diabetes for ESKD may be underestimated. Other issues and limitation of ANZDATA are considered in Chapter 5. Nevertheless, ANZDATA is considered to be of high quality and is a register of all people in Australia receiving dialysis or kidney transplant, where the intention to treat is long-term survival (AIHW 2012a).
Diabetes and chronic kidney disease as risks for other diseases 49
Diabetes and cancer The literature indicates there is a strong association between diabetes (mainly Type 2) and certain types of cancer. Results of several studies have been combined for meta-analysis and have shown that some cancers develop more commonly in people with Type 2 diabetes, including cancer of the liver (Wang et al. 2012), pancreas (Huxley et al. 2005), kidney (Larsson and Wolk 2011), endometrium (Friberg et al. 2007), colon and rectum (Guraya 2015), bladder (Xu et al. 2013) and breast (Boyle et al. 2012). Diabetes has been associated with a decreased risk of prostate cancer (Suh and Kim 2011). This possible protective effect was not included in our study because only linked diseases with increased risk were considered. There have been limited studies on the risk of cancer among people with type 1 diabetes. The exact mechanism of this increased risk of cancer among people with diabetes is unclear, but it is thought that hyperglycaemia and the hyperinsulinemia characteristic of diabetes promotes the growth of cancerous cells, as well as inhibiting the death of these cells (Suh and Kim 2011). Other possible mechanisms include diabetes as a marker of underlying biological factors that alter cancer risk (for example, insulin resistance and hyperinsulinemia). It is also possible, for type 2 diabetes, that the cancer–diabetes association is indirect and due to common risk factors such as obesity and physical inactivity. However, in the selected meta-analyses, these confounders were adjusted in their pooled analysis of effect sizes. Several meta-analyses indicate that the strongest association between type 2 diabetes and cancer risk is with pancreatic and liver cancer (Vigneri et al. 2009); that is, the 2 organs involved in the metabolic problems typical of diabetes.
Diabetes and liver cancer Meta-analyses showed consistent evidence that diabetes is associated with increased risk of liver cancer incidence (El-Serag et al. 2006; Wang et al. 2012, 2014; Yang et al. 2011). Covariates were adjusted; that is, the association found was independent of alcohol use, smoking, hepatitis, liver cirrhosis and high body mass (El-Serag et al. 2006; Wang et al. 2012). It is noted that more than 80% of liver cancer incidence develop in Asian and African countries, and the incidence is relatively low in Australia, USA and Europe (Yang et al. 2011). This is in part explained by the geographical variations in the prevalence of chronic infection with the hepatitis B or hepatitis C viruses: the incidence rate is higher in Eastern Asia and sub-Saharan Africa (Yang et al. 2011). The relative risk used was drawn from Wang et al. (2012), which was assessed as the most up-to-date meta-analysis, including the largest number of studies with an outcome of liver cancer. A total of 17 case-control studies and 32 cohort studies were included in the meta-analysis. The pooled relative risk was 2.31 (95% CI 1.87–2.84) for the association between diabetes and liver cancer with adjusted covariates as described. This was compared with the effect size of 2.50 reported in an earlier meta-analysis by El-Serag et al. (2006) and relative risk of 1.87 reported by Yang et al. (2011).
Diabetes and pancreatic cancer Findings in meta-analyses show increased relative risks of pancreatic cancer in individuals with chronic diabetes and support a modest causal relationship between diabetes and pancreatic cancer (Huxley et al. 2005; Song et al. 2015). One study found that individuals with recently diagnosed diabetes (<4 years) had a 50% greater risk of pancreatic cancer
50 Diabetes and chronic kidney disease as risks for other diseases compared with those who had chronic diabetes (≥5 years) (Huxley et al. 2005). This was consistent with the finding in Song et al. (2015) suggesting diabetes may be an early indicator of pancreatic cancer. There is a study by Wang et al. (2003), however, which suggests that pancreatic cancer can induce diabetes. Patients with pancreatic cancer may have peptides (molecules) that are suggested to have the potential to cause diabetes (Wang et al. 2003). From this inconsistent finding, this study suggests a probable level of evidence of a causal association between diabetes and pancreatic cancer. The odds ratio (1.82; 95% CI 1.66–1.89) for the association between diabetes and pancreatic cancer was drawn from Huxley et al. (2005). This is similar to the relative risk of 1.64 reported by Song et al. (2015) for individuals with duration of diabetes ≥2 years. Song et al. (2015) did not provide an overall relative risk independent of the duration of diabetes. Other meta-analyses found that diabetes was associated with higher mortality in patients with pancreatic cancer (Mao et al. 2015).
Diabetes and bowel cancer There is probable evidence of a causal association between diabetes and bowel cancer as there are some inconsistent findings. Three meta-analyses support a relationship between diabetes and increased risk of bowel cancer (Deng et al. 2012; Guraya 2015; Larsson et al. 2005). Further, a retrospective cohort study by Yang et al. (2004) showed long-term insulin therapy is associated with an increased risk of bowel cancer among patients with type 2 diabetes. This was also supported by Deng et al. (2012). Because it was inferred that insulin use is linked to severity of type 2 diabetes, it is possible that the observed association was due to the severity of diabetes rather than a true effect of exogenous insulin; whether insulin therapy increases bowel cancer risk needs further investigation. Despite this, given the number of meta-analyses available, it was deemed sufficient to include bowel cancer as a linked disease to diabetes for this analysis. The relative risk used was drawn from a meta-analysis by Larsson et al. (2005). This study is older than other meta-analysis available, but it provided details of the studies chosen and the studies were mostly from the USA, Europe and Australia. It included 6 case-control and 9 cohort studies. The pooled relative risk from Larsson et al. (2005) for the association between diabetes and bowel cancer was 1.30 (95% CI 1.20–1.40). It was comparable to other meta- analyses: Deng et al. (2012) reported a relative risk of 1.26 and Guraya (2015) found a relative risk of 1.21, for the association of diabetes and bowel cancer incidence.
Diabetes and uterine and breast cancer Both uterine and breast cancer risks are increased in women with diabetes, and this risk is independent from obesity. Several biological mechanisms have been proposed, most regarding sex hormone abnormalities (Vigneri et al. 2009). High levels of insulin and oestrogens have been suggested to increase endometrial cancer risk by stimulating proliferation of endometrial cells (Friberg et al. 2007).
Diabetes and uterine cancer A meta-analysis by Friberg et al. (2007) found consistent evidence in both case-control and cohort studies on the increased risk of uterine cancer in women with diabetes (largely Type 2) compared with those without diabetes. Studies were conducted in the USA (8), Europe (6), South America (1) and Asia (1). Their pooled relative risk was 2.10 (95% CI 1.75–2.53) for the
Diabetes and chronic kidney disease as risks for other diseases 51
association between diabetes and uterine cancer used in this study. There was no other meta-analysis found in the literature.
Diabetes and breast cancer There is convincing evidence that diabetes is associated with an increased risk of breast cancer. However, studies found this risk in post-menopausal women and not in pre- menopausal women (Boyle et al. 2012; Larsson et al. 2007). To account for this, the analysis used exposure data for women aged over 45. The relative risk used was drawn from Larsson et al. (2007). It included 5 case-control and 15 cohort studies conducted mostly in North America and Europe. The association between diabetes and breast cancer was consistent for these selected studies. It supports a positive association between diabetes and breast cancer risk. The pooled relative risk was 1.20 (95% CI 1.12–1.28), for the association between diabetes and breast cancer in women. This is consistent with a relative risk of 1.27 reported by Boyle et al. (2012). The relative risk from Larsson et al. (2007) was selected as their meta-analysis consisted of a greater proportion of cohort studies (rather than case-control studies) than Boyle et al. (2012).
Diabetes and kidney cancer There is convincing evidence of an association between diabetes and risk of kidney cancer. Meta-analyses found consistent positive association between these conditions (Bao et al. 2013; Larsson & Wolk 2011). Diabetes may affect the risk of kidney cancer by increasing insulin resistance and levels of insulin in the blood that could affect tumour growth (Larsson & Wolk 2011). The relative risk used for this report was from Bao et al. (2013). This large meta-analysis provided details of 18 studies conducted in North America (8), Europe (6) and Asia (4), compared with a higher proportion of studies conducted in Asia used in the meta-analysis by Larsson & Wolk (2011). The pooled relative risk used in this study was 1.40 (95% CI 1.16–1.69) for the association between diabetes and kidney cancer (Bao et al. 2013). This was compared to the relative risk of 1.42 reported by Larsson & Wolk (2011).
Diabetes and bladder cancer There is a strong causal association between diabetes and bladder cancer. This is supported by meta-analyses (Larsson et al. 2006; Xu et al. 2013; Zhu et al. 2013), cohort and case-control studies (Attner et al. 2012; Lo et al. 2013; MacKenzie et al. 2011). The relative risk selected is from a meta-analysis by Larsson et al. (2006); this was assessed to be most applicable for Australian prevalence estimates. The selected studies in this meta- analysis were conducted mostly in North America and Europe, where other meta-analyses included more studies from Asia (Xu et al. 2013; Zhu et al. 2013). It is thought that dietary intake may influence the onset of diabetes, therefore relative risks from countries with a predominantly western diet were selected as being comparable to Australia. Larsson et al. (2006) included 16 studies with a pooled relative risk of 1.24 (95% CI 1.08–1.42), for the association between diabetes and bladder cancer. These results were also similar to other meta-analyses: Xu et al. (2013) reported relative risk of 1.11 and Zhu et al. (2013) found relative risk of 1.35, on the association of diabetes and bladder cancer incidence.
52 Diabetes and chronic kidney disease as risks for other diseases
Table A2: Source of effect size used in calculation of attributable burden for diabetes
Linked Reference Study Number Study Sample Age at Length of Diabetes Outcome Effect size Notes disease type of date size exposure follow-up definition definition (95% CI) studies
Coronary Emerging Risk Meta- 102 Up to May 698,782 Mean = More than Self-report, First ever Male HR Adjusted for age, heart Factor analysis studies 2010 52 years 1 year medication usage myocardial 1.89 (1.73– smoking status, body disease Collaboration and/or baseline infarction or 2.06) mass index and systolic 2010 fasting glucose fatal coronary Females HR blood pressure and, concentration heart disease 2.59 (2.29– where appropriate, ≥7 mmol/L 2.93) stratified by sex and trial arm. Includes both fatal and non-fatal events.
Stroke Emerging Risk Meta- 102 Up to May 698,782 Mean = More than Self-report, Incident Males HR Adjusted for age, Factor analysis studies 2010 52 years 1 year medication usage ischaemic 2.16 (1.84– smoking status, body Collaboration and/or baseline stroke 2.52) mass index and systolic 2010 fasting glucose Females HR blood pressure and, concentration 2.83 (2.35– where appropriate, ≥7 mmol/L 3.40) stratified by sex and trial arm. Includes both fatal and non-fatal events.
Peripheral Fowkes et al. Meta- 34 1997– 112,027 Range = Not noted Fasting glucose Ankle brachial OR 1.88 Odd ratios were based on vascular 2013 analysis studies 2011 25–104 level >7 mmol/L, index ≤0.90 (1.66–2.14) multivariate study design disease years diabetes in which similar medication or definitions of risk factors doctor’s diagnosis (e.g. body mass index, hypertension, smoking and high cholesterol) were used.
Dementia Prince et al. Meta- 11 Up to 35,342 Range = 2–13 Diagnosed and All-cause RR 1.50 Adjusted for age, sex, 2014 analysis studies January 60–88 years undiagnosed dementia (1.33–1.70) education and other 2012 years diabetes (clinically potential risk factors. diagnosed)
(Continued)
Diabetes and chronic kidney disease as risks for other diseases 53
Table A2 (continued): Source of effect size used in calculation of attributable burden for diabetes
Linked Reference Study Number Study Sample Age at Length of Diabetes Outcome Effect size Notes disease type of date size exposure follow-up definition definition (95% CI) studies
Liver cancer Wang et al. Meta- 49 Up to Not Range = 3–15 Self-report, Liver cancer RR 2.31 Adjusted risk estimates 2012 analysis studies February reported 17–80 years medical records confirmed by (1.87–2.84) were used. Studies 2011 years or database histological and adjusted for age, sex, review, blood cytological hepatitis, alcohol, glucose level or examinations, smoking, body mass medications laboratory tests index and liver or imaging cirrhosis. findings
Pancreatic Huxley et al. Meta- 36 Up to 9,220 Not Not Self-report, Pancreatic OR 1.82 Adjusted for sex, age, cancer 2005 analysis studies January reported reported medical records cancer (1.66–1.89) smoking, social class 2005 or blood glucose confirmed by and dietary risk factors. level laboratory tests, imaging findings, cancer registry
Bowel Larsson et al. Meta- 15 Up to July 2,593,935 Not Not Self-report, Incident bowel RR 1.30 Adjusted for age, sex, cancer 2005 analysis studies 2005 reported reported medical records cancer diagnosis (1.20–1.40) body mass index, or fasting smoking, alcohol, glucose level ≥7 hormone replacement mmol/L therapy use and dietary risk factors.
Breast Larsson et al. Meta- 20 Up to 1,400,000 Range = Not Self-report, Incident breast RR 1.20 Adjusted for age, body cancer 2007 analysis studies February 20–95 reported medical records cancer diagnosis (1.12–1.28) mass index, smoking, 2007 years or blood glucose alcohol, social class level and dietary risk factors.
Uterine Friberg et al. Meta- 16 Up to 96,003 ≥20 years Not Self-report, Incident uterine RR 2.10 Adjusted for age, body cancer 2007 analysis studies January reported medical records cancer diagnosis (1.75–2.53) mass index, smoking, 2007 or medical alcohol, hypertension, registers education, physical activity, parity and dietary risk factors.
(Continued)
54 Diabetes and chronic kidney disease as risks for other diseases
Table A2 (continued): Source of effect size used in calculation of attributable burden for diabetes
Linked Reference Study Number Study Sample Age at Length of Diabetes Outcome Effect size Notes disease type of date size exposure follow-up definition definition (95% CI) studies
Kidney Bao et al. Meta- 24 studies Up to 6,025,827 Not 6–40 years Self-report or Incident kidney RR 1.40 Adjusted for age, sex, cancer 2013 analysis February reported medical records cancer (1.16–1.69) ethnicity, body mass 2012 diagnosis index, smoking, alcohol, physical activity and dietary risk factors.
Bladder Larsson et Meta- 16 studies Up to July Not Not Not Self-report, Incident bladder RR 1.24 Adjusted for age, sex, cancer al. 2006 analysis 2006 reported reported reported medical records or cancer (1.08–1.42) smoking, body mass fasting glucose diagnosis index and dietary risk level factors.
Diabetes and chronic kidney disease as risks for other diseases 55
Appendix B: Selection of effect sizes for chronic kidney disease
A review of the relevant literature identified that if an individual has CKD, there is an increased risk of developing a number of diseases: 4 diseases were identified for this study. In this appendix, effect size measures the risk of developing the specific linked disease among those with and without CKD. The effect sizes used in this study to estimate the indirect burden of CKD and their sources are listed in Table B1. For further information on the selection of linked diseases and effect sizes see Chapter 2. A summary of selected studies is presented in Table B2. CKD in this appendix mainly refers to stages 3 to 5, defined in the literature by eGFR of less than 60 mL/min/1.73 m2 or the levels of urinary protein or albumin. The definition of eGFR is consistent with stage 3 to 5 of CKD used in the ABDS 2011.
Table B1: Effect size and sources for chronic kidney disease and linked disease pair analysis
Linked disease Level of evidence Effect size (95% CI) Source of effect size
Coronary heart disease Convincing RR 1.47 (1.23–1.74) Perkovic et al. 2008 Stroke (all types) Convincing RR 1.43 (1.31–1.57) Lee et al. 2010a Peripheral vascular disease Probable OR 2.5 (1.2–5.1) O'Hare et al. 2003 Dementia (all types; age 65+) Possible/probable RR 1.37 (1.06–1.78) Seliger et al. 2004
CKD and cardiovascular disease There is strong evidence to suggest that CKD is an independent risk factor for cardiovascular disease (Herzog et al. 2011; Sarnack et al. 2003). Many studies have shown markers of kidney dysfunction such as raised proteinuria (urinary protein excretion) or albuminuria (urinary albumin excretion), and low estimated glomerular filtration rate (eGFR) are associated with cardiovascular disease (Perkovic et al. 2008). Morbidity and mortality from cardiovascular disease are inversely and independently associated with kidney function; the incidence and severity of cardiovascular disease increases as eGFR declines (Gansevoort et al. 2013). Meta-analyses also show the risk of cardiovascular mortality increases linearly with decreasing eGFR, after adjustment of cardiovascular risk factors (Matsushita et al. 2010; Van der Velde et al. 2011). Mechanisms specific to CKD promote vascular disease. Patients with CKD have higher risk of multi-vessel coronary calcification (Herzog et al. 2011) and in patients with advancing CKD, the prevalence of left-ventricular hypertrophy is increased, which leads to reduced coronary reserve (Gansevoort et al. 2013). Other studies have shown that CKD promotes hypertension and dyslipidaemia (Kokubo et al. 2009; Schiffrin et al. 2007), which could accelerate atherosclerosis and lead to increased prevalence of coronary artery disease, heart failure, stroke and peripheral vascular disease.
CKD and coronary heart disease There is strong and consistent evidence for an association between proteinuria and albuminuria with subsequent risk of coronary heart disease. Cohort studies have shown that microalbuminuria is a strong and independent determinant of coronary heart disease and
56 Diabetes and chronic kidney disease as risks for other diseases death (Klausen et al. 2004; Ryoo et al. 2011). Further, CKD is associated with risk for recurrent coronary heart disease that is similar to other high-risk conditions, such as diabetes, metabolic syndrome or cigarette smoking (Baber et al. 2013). The pooled relative risk used (1.47, 95% CI 1.23–1.74) for the association between CKD and coronary heart disease was sourced from a meta-analysis by Perkovic et al. (2008). The results from this meta-analysis of 26 cohort studies, including information on over 7,000 coronary heart disease events among almost 170,000 individuals, suggested that people with proteinuria have a risk of coronary heart disease that is 50% greater than those without. Furthermore, there was evidence to indicate a dose–response relationship where the strength of the association was higher among individuals with macroalbuminuria compared with those with microalbuminuria. The relationship was consistent across predefined population subgroups, including sex, ethnicity and individuals with and without diabetes. This meta-analysis confirmed a strong and independent association between chronic kidney dysfunction and risk of coronary heart disease.
CKD and stroke There is strong and consistent evidence for an association between CKD and stroke. Several meta-analyses found the risk of stroke increases linearly with decreasing eGFR or increasing albuminuria (Lee et al. 2010a, 2010b ; Masson et al. 2015). People with a low eGFR (<60 mL/min/1.73 m2) had an independent risk of stroke greater than those with a normal baseline eGFR (Lee et al. 2010a; Weiner et al. 2004). The pooled relative risk used (1.43, 95% CI 1.31–1.57) for the association between CKD and stroke was sourced from a meta-analysis by Lee et al. (2010a). This meta-analysis was assessed to be the most up to date, including the largest number of studies with an outcome of any stroke. It included 21 articles derived from 33 prospective studies and over 280,000 persons experiencing almost 8,000 stroke events. The results found a link between low eGFR and risk of future stroke. This relationship was consistent across diverse population subgroups—that is, those with or without cardiovascular risk factors. The pooled relative risk of 1.43 for the risk of stroke was similar to the hazard ratio of 1.17 reported by Weiner et al. (2004), where CKD was also defined by low eGFR. In contrast, the relative risk reported by Lee et al. (2010b) for macroalbuminuria was 2.65 and for microalbuminuria was 1.58, while the relative risk for proteinuria was 1.71 reported in a meta-analysis by Ninomiya et al. (2009).
CKD and peripheral vascular disease There are few high-quality observational studies on CKD and the risk of peripheral vascular disease. The cohort study by O’Hare et al. (2004) found a strong association between renal insufficiency and lower-extremity peripheral vascular disease, independent of potential confounders such as age, diabetes, hypertension, coronary heart disease, stroke history and high cholesterol. The odds ratio was 2.5 (95% CI 1.2–5.1). This study was selected because it had a large sample size of 2,229 persons and clearly defined methodology. The level of evidence for CKD was assessed to be probable in this report, because evidence was obtained from observational studies.
Diabetes and chronic kidney disease as risks for other diseases 57
CKD and dementia The level of evidence for CKD was assessed to be possible/probable in this report. This is because there are limited studies demonstrating a causal association between CKD and clinically defined dementia (Helmer et al. 2011; Miwa et al. 2014; Sasaki et al. 2011; Seliger et al. 2004). A systematic review of the evidence found the association to be consistent between studies for CKD as a risk factor for cognitive decline, but no meta-analyses have been conducted (Deckers et al. 2015). Cognitive decline is less severe than dementia and not all persons with cognitive decline go on to have dementia, with a dementia diagnosis requiring severe cognitive decline and meeting diagnostic criteria. Three studies have reported that impaired kidney function is associated with higher risk of dementia after adjusting for confounding factors: brain atrophy and small blood vessel disease (Miwa et al. 2014; Sasaki et al. 2011; Seliger et al. 2004). Another study showed no association between low kidney function at baseline and dementia, but found that a fast decline in kidney function—probably reflecting more severe kidney damage or impairment—was associated with an increased risk of dementia (Helmer et al. 2011). The relative risk of 1.37 used in this project is taken from a 2004 study of 3,349 persons aged over 65 undertaken in the USA (Seliger et al. 2004). Moderate renal impairment was defined as elevated serum creatinine levels according to gender specific cut-offs (≥1.3 mg/dL for women and ≥1.5 mg/dL for men) (Seliger et al. 2004). Based on available evidence, this was the most applicable measure of CKD for Australian prevalence estimates. It was also the most comparable study for an Australian context, with other prospective longitudinal studies undertaken in Japanese and South-East Asian populations (Miwa et al. 2014; Sasaki et al. 2011). Several biologically plausible mechanisms may explain how impaired kidney function leads to increased risk of dementia (Bugnicourt et al. 2013). Among these are anaemia, atherosclerosis and increased risk of stroke and micro infarcts, as well as elevated homocysteine—all more common in people with CKD and associated with increased dementia risk (Bugnicourt et al. 2013; Etgen 2015).
58 Diabetes and chronic kidney disease as risks for other diseases
Table B2: Source of effect size used in calculation of attributable burden for chronic kidney disease
Linked Reference Study type Number Study Sample Age at Length of CKD definition Outcome Effect size Notes disease of date size exposure follow-up definition (95%CI) studies
Coronary Perkovic et Meta- 26 Up to 169,949 >20 years 4–27 Proteinuria or Fatal or non- RR 1.47 Adjusted for age, sex, heart al. 2008 analysis cohort November years albuminuria fatal coronary (1.23–1.74) body mass index, disease studies 2006 heart disease smoking, alcohol, hypertension, diabetes, education, total cholesterol and physical activity.
Stroke Lee et al. Meta- 21 Up to 284,672 Range = 3–15 eGFR All fatal and RR 1.43 Adjusted for age, sex, 2010a analysis studies October 53–78 years <60 mL/min/1.73 m2 non-fatal (1.31–1.57) body mass index, 2009 years stroke smoking, alcohol, (ischaemic and hypertension, diabetes, haemorrhagic) education, coronary heart disease, diabetes, cholesterol and selected medications.
Peripheral O'Hare et Prospective 1 study 1999– 2,229 ≥40 years Not eGFR Ankle brachial OR 2.5 Adjusted for age, sex, vascular al. 2003 population 2000 reported <60 mL/min/1.73 m2 index <0.90 (1.2–5.1) body mass index, disease study ethnicity, diabetes, hypertension, coronary artery disease, stroke history.
Dementia Seliger et Prospective 1 study 1989– 3,349 Mean = 76 Median 6 Elevated serum All-cause RR 1.37 Adjusted for age, sex, al. 2004 longitudinal 1990 ± 4.9 years creatinine dementia (1.06–1.78) education, ethnicity, population and years Males ≥1.5 mg/dL (clinically body mass index, study Females diagnosed coronary heart disease, 1992– DSM-IV(a)) hypertension, diabetes, 1999 ≥1.3 mg/dL smoking, apolipoprotein (apoE) genotype.
Diabetes and chronic kidney disease as risks for other diseases 59
Appendix C: Methods for scenario modelling
In this report, the collective burden due to diabetes and CKD was extrapolated to 2020 using 2 scenarios. The difference between these scenarios provide an indication of the amount of burden that may be avoided if the rise of diabetes or CKD is halted, compared with the amount of burden if the current trends continue. Results from both methods were compared with the 2011 burden to estimate the extent of impact between the 2 scenarios (chapters 3 and 4). The year 2020 was chosen because it aligns with the National Strategic Framework for Chronic Conditions and the World Health Organization’s Global Action Plan for the Prevention and Control of Non-communicable Diseases 2013–2020 (WHO 2013).
Indirect burden Projected indirect burden is dependent on both the future exposure of the risk factor (in this case exposure to diabetes and CKD) and the future burden of the linked diseases, which may be influenced by many factors such as diagnosis and treatment, as well as exposure to other risk factors.
Diabetes and CKD exposure To estimate the indirect burden of diabetes and CKD in 2020, diabetes and CKD exposure (in this case, disease prevalence) were extrapolated using 2 scenarios. Scenario A assumes the past trend of prevalence rate (between 2003 and 2011) continues to 2020, while scenario B assumes that the 2011 prevalence rate will remain the same to 2020. Both scenarios account for expected changes in the population structure. The 2 scenarios used to extrapolate diabetes and CKD prevalence in 2020 used the following methods: • Scenario A The annual rate of change between the prevalence rate in 2003 and 2011 from the ABDS 2011 and extrapolated to 2020, was applied to projected 2020 populations. Other information about past trends in these conditions was used to determine if the rate of change is likely to continue. Assumption: the annual prevalence rate of change between 2003 and 2011 will continue to 2020. • Scenario B Applying the prevalence rate in 2011 to projected 2020 populations; obtained from the Australian Bureau of Statistics population projections series B (ABS 2013). Assumption: the prevalence rate in 2011 is the same in 2020. The only change in burden is due to population growth and ageing. The results show the differences in attributable burden that could be expected if these scenarios were to reflect the population exposure.
60 Diabetes and chronic kidney disease as risks for other diseases
The data sources and methods used to derive prevalence for diabetes and CKD in 2011 and 2003 are described in Appendix D. The annual rate of change in prevalence and fatal burden between 2003 and 2011 was derived for each sex and age group (see tables C1–C4). The annual rate of change in non-fatal burden between 2003 and 2011 were calculated at the sequela level for each sex and age group. Sequelae are consequences associated with the disease (see Appendix D) and extrapolation of trends at the sequelae level provides more accurate estimates of non-fatal burden.
Linked diseases For both scenarios, it was assumed the 2011 burden rates of the linked diseases will remain the same to 2020, adjusted for expected changes in the population structure. Due to the complexity of possible associations between diseases, expected future changes in linked disease burden will require more consideration. This assumption was made for simplicity in our analysis. The indirect diabetes and CKD burden estimated in 2020 under different prevalence scenarios for each linked disease are summarised in tables C5 and C6.
Direct burden The direct burden for diabetes and CKD was also extrapolated to 2020 using similar scenarios. Scenario A assumes that the past trend of fatal and non-fatal burden rate (between 2003 and 2011) continues to 2020, while scenario B assumes the 2011 fatal and non-fatal burden rate will remain the same in 2020, where the only change in burden is due to population growth and ageing. Both scenarios account for expected changes in the population structure.
Collective burden For each scenario A and B, the sum of indirect and direct burden provides estimates of the collective burden in 2020.
Assumptions and limitations In our scenario analysis, we extrapolated diabetes and CKD exposure to 2020 and examined their impact on linked diseases. These extrapolations are, by nature, estimates about what might reasonably be expected in the future. Estimates based on historical trends and other available information can generate the best estimates. However, there is no guarantee of their realisation in the future. Scenario A and B generated different estimates of attributable burden in 2020. A major limitation for scenario A is that the assumption on the rate of change was based on 2 time points (2003 and 2011). Updated burden of disease estimates may provide opportunity for future extrapolations based on more than 2 time points, leading to increased mathematical accuracy. A limitation is that, when using scenarios to estimate the impact of reduced exposure to diabetes or CKD on linked disease burden, it is not possible to quantify the flow-on effects to other possible diseases in the causal pathway that may occur in reality. For example, estimating the impact of reducing exposure to diabetes on coronary heart disease burden is
Diabetes and chronic kidney disease as risks for other diseases 61 independent of measuring the impact of any potential reduction in CKD that might arise from reducing exposure to diabetes. In this study, the impact of diabetes and CKD on linked diseases was assessed independently. Similarly, due to the complexity of possible associations between diseases, expected future changes in linked disease burden will require more consideration. It was assumed the 2011 burden rates of the linked diseases will remain constant to 2020. This assumption was made for simplicity in our analysis.
62 Diabetes and chronic kidney disease as risks for other diseases
Table C1: Estimated diabetes prevalence rates (number per 1,000) for 2003 and 2011 used in scenario modelling
Rate change between Rate change between 2003 2011 Annual rate difference 2003 and 2011 2011 and 2020 Age group (years) Males Females Males Females Males Females Males Females Males Females
0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1–4 0.5 0.4 0.4 0.3 0.0 0.0 –0.1 0.0 –0.1 0.0 5–9 1.6 1.6 1.3 1.4 0.0 0.0 –0.3 –0.3 –0.4 –0.3
10–14 3.7 3.6 2.9 2.9 –0.1 –0.1 –0.8 –0.7 –0.9 –0.7 15–19 3.0 3.0 2.5 2.5 –0.1 –0.1 –0.5 –0.4 –0.6 –0.5 20–24 2.8 1.3 2.6 1.3 0.0 0.0 –0.1 0.0 –0.2 0.0
25–29 9.9 3.9 9.7 3.9 0.0 0.0 –0.2 0.0 –0.2 0.0 30–34 19.5 8.9 15.8 7.4 –0.5 –0.2 –3.7 –1.5 –4.1 –1.7 35–39 22.5 13.7 23.1 23.9 0.1 1.3 0.6 10.3 0.7 11.5
40–44 30.1 19.6 42.0 31.2 1.5 1.4 12.0 11.6 13.4 13.0 45–49 40.4 25.1 63.7 36.6 2.9 1.4 23.3 11.5 26.2 12.9 50–54 54.2 31.2 90.0 37.0 4.5 0.7 35.8 5.9 40.3 6.6
55–59 100.7 53.4 107.9 85.8 0.9 4.1 7.2 32.4 8.1 36.5 60–64 125.1 67.6 177.2 93.5 6.5 3.2 52.1 25.8 58.6 29.1 65–69 125.5 72.6 245.3 127.2 15.0 6.8 119.8 54.6 134.8 61.4
70–74 148.0 90.0 225.5 113.5 9.7 2.9 77.5 23.4 87.2 26.4 75–79 158.4 99.5 190.2 134.0 4.0 4.3 31.8 34.5 35.7 38.8 80–84 126.9 85.7 211.4 104.0 10.6 2.3 84.5 18.3 95.0 20.6
85–89 119.3 85.3 184.0 117.4 8.1 4.0 64.7 32.2 72.8 36.2 90–94 141.7 108.8 215.9 148.0 9.3 4.9 74.1 39.2 83.4 44.1 95–99 122.7 160.6 183.0 235.7 7.5 9.4 60.3 75.1 67.8 84.5
100+ 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Source: AIHW burden of disease database, 2011.
Diabetes and chronic kidney disease as risks for other diseases 63
Table C2: Diabetes YLL rates (number per 1,000) for 2003 and 2011 used in scenario modelling
Rate change between Rate change between 2003 2011 Annual rate change 2003 and 2011 2011 and 2020 Age group (years) Males Females Males Females Males Females Males Females Males Females
0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1–4 0.0 0.2 0.0 0.0 0.0 0.0 0.0 –0.2 0.0 –0.2 5–9 0.0 0.0 0.2 0.0 0.0 0.0 0.2 0.0 0.2 0.0
10–14 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.1 0.0 0.1 15–19 0.2 0.2 0.1 0.2 0.0 0.0 –0.1 0.0 –0.1 0.0 20–24 0.1 0.1 0.2 0.2 0.0 0.0 0.1 0.1 0.1 0.1
25–29 0.3 0.2 0.2 0.1 0.0 0.0 –0.1 0.0 –0.2 0.0 30–34 0.8 0.4 0.2 0.4 –0.1 0.0 –0.6 0.0 –0.7 0.0 35–39 0.8 0.2 0.6 0.1 0.0 0.0 –0.2 –0.1 –0.2 –0.1
40–44 1.3 0.6 1.0 0.6 0.0 0.0 –0.3 0.0 –0.3 0.0 45–49 1.1 1.2 1.4 1.1 0.0 0.0 0.3 –0.1 0.4 –0.2 50–54 3.1 1.1 2.7 1.6 –0.1 0.1 –0.4 0.5 –0.5 0.5
55–59 4.6 1.5 3.0 1.7 –0.2 0.0 –1.6 0.1 –1.8 0.2 60–64 7.6 3.6 5.3 2.8 –0.3 –0.1 –2.3 –0.8 –2.6 –0.9 65–69 11.8 6.1 9.6 4.7 –0.3 –0.2 –2.2 –1.4 –2.5 –1.6
70–74 17.5 8.4 11.7 7.2 –0.7 –0.1 –5.8 –1.2 –6.5 –1.3 75–79 17.3 12.8 17.4 10.1 0.0 –0.3 0.0 –2.7 0.0 –3.1 80–84 21.1 13.9 22.3 14.1 0.2 0.0 1.2 0.2 1.4 0.2
85–89 27.0 18.5 24.0 19.9 –0.4 0.2 –3.0 1.4 –3.4 1.5 90–94 26.6 17.9 27.0 20.0 0.1 0.3 0.4 2.0 0.5 2.3 95–99 16.3 18.0 24.7 20.7 1.1 0.3 8.4 2.7 9.5 3.1
100+ 11.3 11.8 25.1 21.6 1.7 1.2 13.9 9.8 15.6 11.0
Source: AIHW burden of disease database, 2011.
64 Diabetes and chronic kidney disease as risks for other diseases
Table C3: Estimated chronic kidney disease prevalence rates (number per 1,000) for 2003 and 2011 used in scenario modelling
Rate change between Rate change between 2003 2011 Annual rate difference 2003 and 2011 2011 and 2020 Age group (years) Males Females Males Females Males Females Males Females Males Females
0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1–4 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5–9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 10–14 0.1 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 15–19 0.3 0.3 0.3 0.4 0.0 0.0 0.1 0.1 0.1 0.1 20–24 0.9 0.8 1.0 0.9 0.0 0.0 0.1 0.1 0.1 0.1 25–29 1.5 1.6 1.6 1.8 0.0 0.0 0.1 0.1 0.2 0.1 30–34 1.6 1.9 2.0 2.4 0.0 0.1 0.4 0.5 0.4 0.6 35–39 2.7 3.0 3.4 3.7 0.1 0.1 0.7 0.7 0.8 0.8 40–44 3.3 3.6 4.5 4.6 0.1 0.1 1.2 1.1 1.3 1.2 45–49 5.0 5.0 6.5 6.1 0.2 0.1 1.5 1.2 1.7 1.3 50–54 6.7 6.4 8.5 7.6 0.2 0.2 1.8 1.3 2.0 1.4 55–59 10.1 9.0 13.0 10.6 0.4 0.2 3.0 1.6 3.3 1.8 60–64 19.1 15.8 20.4 15.3 0.2 –0.1 1.3 –0.5 1.5 –0.6 65–69 83.3 61.9 91.5 64.2 1.0 0.3 8.2 2.3 9.2 2.6
70–74 133.4 92.1 165.1 111.5 4.0 2.4 31.7 19.3 35.7 21.8 75–79 129.8 145.7 176.3 197.7 5.8 6.5 46.5 52.0 52.3 58.5 80–84 255.8 276.2 286.1 329.5 3.8 6.7 30.3 53.4 34.1 60.0 85–89 285.8 302.7 264.4 305.2 –2.7 0.3 –21.4 2.4 –24.1 2.7 90–94 333.6 337.7 312.7 344.5 –2.6 0.8 –20.9 6.8 –23.6 7.6 95–99 320.4 376.9 308.0 357.7 –1.5 –2.4 –12.4 –19.2 –13.9 –21.6
100+ 167.0 262.2 244.2 255.5 9.7 –0.8 77.2 –6.6 86.9 –7.5
Source: AIHW burden of disease database, 2011.
Diabetes and chronic kidney disease as risks for other diseases 65
Table C4: Chronic kidney disease YLL rates (number per 1,000) for 2003 and 2011 used in scenario modelling
Rate change between Rate change between 2003 2011 Annual rate change 2003 and 2011 2011 and 2020 Age group (years) Males Females Males Females Males Females Males Females Males Females
0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1–4 0.0 0.2 0.0 0.0 0.0 0.0 0.0 –0.2 0.0 –0.2 5–9 0.0 0.0 0.2 0.0 0.0 0.0 0.2 0.0 0.2 0.0
10–14 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.1 0.0 0.1 15–19 0.2 0.2 0.1 0.2 0.0 0.0 –0.1 0.0 –0.1 0.0 20–24 0.1 0.1 0.2 0.2 0.0 0.0 0.1 0.1 0.1 0.1
25–29 0.3 0.2 0.2 0.1 0.0 0.0 –0.1 0.0 –0.2 0.0 30–34 0.8 0.4 0.2 0.4 –0.1 0.0 –0.6 0.0 –0.7 0.0 35–39 0.8 0.2 0.6 0.1 0.0 0.0 –0.2 –0.1 –0.2 –0.1
40–44 1.3 0.6 1.0 0.6 0.0 0.0 –0.3 0.0 –0.3 0.0 45–49 1.1 1.2 1.4 1.1 0.0 0.0 0.3 –0.1 0.4 –0.2 50–54 3.1 1.1 2.7 1.6 –0.1 0.1 –0.4 0.5 –0.5 0.5
55–59 4.6 1.5 3.0 1.7 –0.2 0.0 –1.6 0.1 –1.8 0.2 60–64 7.6 3.6 5.3 2.8 –0.3 –0.1 –2.3 –0.8 –2.6 –0.9 65–69 11.8 6.1 9.6 4.7 –0.3 –0.2 –2.2 –1.4 –2.5 –1.6
70–74 17.5 8.4 11.7 7.2 –0.7 –0.1 –5.8 –1.2 –6.5 –1.3 75–79 17.3 12.8 17.4 10.1 0.0 –0.3 0.0 –2.7 0.0 –3.1 80–84 21.1 13.9 22.3 14.1 0.2 0.0 1.2 0.2 1.4 0.2
85–89 27.0 18.5 24.0 19.9 –0.4 0.2 –3.0 1.4 –3.4 1.5 90–94 26.6 17.9 27.0 20.0 0.1 0.3 0.4 2.0 0.5 2.3 95–99 16.3 18.0 24.7 20.7 1.1 0.3 8.4 2.7 9.5 3.1
100+ 11.3 11.8 25.1 21.6 1.7 1.2 13.9 9.8 15.6 11.0
Source: AIHW burden of disease database, 2011.
66 Diabetes and chronic kidney disease as risks for other diseases
Table C5: Indirect diabetes burden estimated in 2020 under different prevalence scenarios, by linked disease Scenario B compared with Scenario A Scenario B Scenario A
Continued trend in diabetes prevalence Stable diabetes prevalence rate to 2020 DALY that would be avoided in 2020 rate to 2020
Percentage change Percentage change Attributable Attributable DALY Attributable DALY (b) Linked disease from 2011 from 2011 DALY number DALY (%) DALY in 2011 in 2020 (a) in 2020 (a) attributable DALY attributable DALY
Coronary heart disease 38,852 79,588 104.8 50,964 31.2 28,624 36.0
Stroke 18,730 37,420 99.8 24,611 31.4 12,808 34.2
Chronic kidney disease 8,945 16,950 89.5 11,547 29.1 5,403 31.9
Dementia 8,018 18,033 124.9 10,849 35.3 7,184 39.8
Liver cancer 3,655 7,395 102.3 4,728 29.4 2,667 36.1
Pancreatic cancer 3,655 7,002 91.6 4,660 27.5 2,342 33.4
Bowel cancer 2,993 6,420 114.5 3,910 30.6 2,510 39.1
Breast cancer 875 1,783 103.8 1,106 26.4 677 38.0
Peripheral vascular 771 1,651 114.1 1,025 33.0 626 37.9 disease
Kidney cancer 739 1,553 110.1 949 28.5 603 38.8
Uterine cancer 610 1,192 95.3 783 28.3 409 34.3
Bladder cancer 489 1,097 124.3 655 33.9 442 40.3
Indirect burden total 88,332 180,084 103.9 115,788 31.1 64,297 35.7
(a) Percentage change is the change in attributable DALY between 2020 and 2011 divided by the attributable DALY in 2011 for each scenario.
(b) Percentage difference is the difference in attributable DALY in 2020 between Scenario A and B divided by the attributable DALY in 2020 for Scenario A.
Source: AIHW analysis of burden of disease database, 2011.
Diabetes and chronic kidney disease as risks for other diseases 67
Table C6: Indirect chronic kidney disease burden under different exposure scenarios, by linked disease Scenario B compared with Scenario A Scenario B Scenario A
Continued trend in CKD prevalence rate Stable CKD prevalence rate to 2020 DALY that would be avoided in 2020 to 2020
Percentage change Percentage change Attributable Attributable DALY Attributable DALY (b) Linked disease from 2011 from 2011 DALY number DALY (%) DALY in 2011 in 2020 (a) in 2020 (a) attributable DALY attributable DALY
Coronary heart disease 22,728 33,134 45.8 30,401 33.8 2,733 8.2
Dementia 12,678 17,868 40.9 16,789 32.4 1,079 6.0
Stroke 9,859 14,097 43.0 12,978 31.6 1,119 7.9
Peripheral vascular 1,621 2,312 42.6 2,173 34.1 139 6.0 disease
Indirect burden total 46,886 67,410 43.8 62,342 33.0 5,068 7.5
(a) Percentage change is the change in attributable DALY between 2020 and 2011 divided by the attributable DALY in 2011 for each scenario.
(b) Percentage difference is the difference in attributable DALY in 2020 between Scenario A and B divided by the attributable DALY in 2020 for Scenario A.
Source: AIHW analysis of burden of disease database, 2011.
68 Diabetes and chronic kidney disease as risks for other diseases
Appendix D: Data sources
Prevalence estimates and direct burden estimates for diabetes and CKD were obtained from the ABDS 2011. Detailed methodology on the methods and assumptions used by the ABDS 2011 can be found in Australian Burden of Disease Study 2011: methods and supplementary material (AIHW 2016b). For diabetes and CKD, prevalence estimates were calculated at the sequela level (consequences associated with diabetes); estimates for each sequela were then combined to produce burden estimates for diabetes and CKD, respectively.
Diabetes Diabetes includes type 1, type 2 and other diabetes types, with the exception of gestational diabetes (included in reproductive and maternal conditions). Two sequelae were included for analysis: undiagnosed and diagnosed diabetes. Prevalence estimates for undiagnosed and diagnosed diabetes in people aged 18 and over were derived from the biomedical component of the Australian Health Survey (AHS) 2011– 12. For those aged under 18, diagnosed diabetes prevalence estimates were obtained from the National Diabetes Registry. It was assumed that there were no people with undiagnosed diabetes aged under 18.
2003 prevalence estimates (used to inform current trends in scenario modelling) For diagnosed diabetes, because the health surveys before the AHS did not have biomedical components, the 2003 prevalence estimates were modelled using the self-report data from the 2001, 2004–05, 2007–08 National Health Survey and the AHS.
Chronic kidney disease The primary data source for estimating prevalence of chronic kidney disease (with and without anaemia) was the AHS 2011–12, while the primary data source for estimating prevalence of ESKD was the Australia and New Zealand Dialysis and Transplant Registry (ANZDATA) 2011. There were 6 sequelae included for analysis: asymptomatic; anaemia due to stage 3 chronic kidney disease; stage 4 chronic kidney disease; anaemia due to stage 4 chronic kidney disease; ESKD treated with dialysis or transplant; and untreated ESKD. Asymptomatic The prevalence of asymptomatic stage 3 CKD was estimated using measured data from the AHS 2011–12. Stages were determined by combining the participants' estimated glomerular filtration rate results with their albumin:creatinine ratio results as described in Cardiovascular disease, diabetes and chronic kidney disease—Australian facts: prevalence and incidence (AIHW 2014). It is important to note that asymptomatic CKD was given an asymptomatic health state, so it did not contribute to the non-fatal burden.
Diabetes and chronic kidney disease as risks for other diseases 69
Anaemia due to stage 3 chronic kidney disease The AHS 2011–12 was used to calculate the proportion of people with stage 3 chronic kidney disease by broad age group and sex. The age and sex distribution was further refined using hospitalisations data from the National Hospital Morbidity Database. Stage 4 chronic kidney disease and Anaemia due to stage 4 chronic kidney disease The prevalence of stage 4 CKD was also estimated using measured data from the AHS 2011–12, using the estimate of persons with stage 4 and 5 chronic kidney disease minus those with ESKD (stage 5 only) sourced from the ANZDATA. The age and sex distribution was based on the AHS 2011–12 results but disaggregated further by the age and sex of persons who were hospitalised in 2011. End stage kidney disease treated with dialysis or transplant Registry data from the 2011 ANZDATA was used to determine the prevalence of ESKD treated by dialysis or transplant. Untreated ESKD Untreated ESKD refers to people were not receiving kidney replacement therapy, although they may be receiving palliative treatments. The prevalence of people with untreated ESKD was estimated from an analysis of the 2010 ANZDATA linked with the AIHW National Mortality Database and National Death Index to identify persons who died from ESKD that were not treated with kidney replacement therapy or were not in the registry (AIHW 2011). The prevalence for 2010 was assumed to be the same as 2011.
2003 prevalence estimates (used to inform current trends in scenario modelling) The ratio of the prevalence of treated ESKD to the prevalence of stage 3 CKD and stage 4 CKD in 2011 was used to estimate prevalence in 2003, due to lack of biomedical measurement data consistent with the 2011 method. The age and sex distribution from 2011 were then applied to these estimates because the hospital codes used to estimate these were not in use in 2003.
Population attributable fraction for CKD attributable to diabetes The proportion of CKD that was attributable to diabetes (population attributable fraction) was calculated based on the ANZDATA, because the ANZDATA only includes data on ESKD receiving treatment (stage 5 CKD). In our analysis, the population attributable fraction was assumed to apply to all stages of CKD. The variable used to calculate the proportion of people with CKD due to diabetes in the ANZDATA is the ‘primary group’—specifically the proportion of individuals with diabetic nephropathy by age and sex. The registry includes limited information on people aged 90 and over because they are less likely to get renal replacement therapy over time. As such, the PAFs were modelled for age 90 and over.
Linked diseases The data sources for the DALY estimates of each linked diseases are summarised in Table D1.
70 Diabetes and chronic kidney disease as risks for other diseases
Table D1: The Australian Burden of Disease Study 2011 main data sources for YLD estimation Linked disease Key national data sources
Cancers Australian Cancer Database
National Mortality Database
National Hospital Morbidity Database
Epidemiological studies
Cardiovascular diseases (coronary National Hospital Morbidity Database heart disease, stroke, peripheral vascular disease) Western Australian linked data
New Zealand Burden of Disease Study
Dementia AIHW dementia analyses International epidemiological studies
Diabetes and chronic kidney disease as risks for other diseases 71
Glossary attributable burden: The disease burden attributed with a particular risk factor. It is the reduction in fatal and non-fatal burden that would have occurred if exposure to the risk factor had been avoided (or, more precisely, had been at its theoretical minimum). chronic: Persistent and long-lasting. collective burden: The sum of the direct and indirect burden. comorbidity: A situation where a person has 2 or more health problems at the same time. comparative risk assessment: The process for estimating the burden of disease attributable to selected risk factors. It involves 5 key steps: selection of risk–outcome pairs; estimation of exposure distribution; estimation of effect sizes; choice of theoretical minimum risk exposure level; and finally the calculation of attributable burden. DALY (disability-adjusted life years): Measure (in years) of healthy life lost, either through premature death defined as dying before the expected life span (YLL) or, equivalently, through living with ill health due to illness or injury (YLD). direct burden: In burden of disease analysis, it is the burden calculated to capture the main disabling consequences of the disease. For example, the direct diabetes burden includes diabetic nephropathy, neuropathy and retinopathy. disability: In burden of disease analysis, any departure from an ideal health state. disease: A broad term that can be applied to any health problem, including symptoms, diseases, injuries and certain risk factors, such as high blood cholesterol and obesity. Often used synonymously with condition, disorder or problem. diseases-as-risks: Diseases act as risk factors for other diseases. To fully account for the health loss attributable to diseases-as-risks requires that their ‘indirect’ burdens be calculated and then added to their ‘direct’ burdens in order to estimate their collective burdens. effect size: A statistical measure of the strength of the relationship between 2 variables (in this context, between a risk exposure and a disease outcome), expressed, for example, as a relative risk. fatal burden: The burden from dying ‘prematurely’ as measured by years of life lost. Often used synonymously with YLL, and also referred to as ‘life lost’. hazard ratio (HR): Hazard ratio has similar interpretation to relative risk. It is the ratio of the probability of an event (e.g. disease outcome) in the exposed group to the probability in the control group. Hazard ratios differ from relative risks in that the latter are cumulative over an entire study, using a defined endpoint, while the former represent instantaneous risk at some particular time period during the study. indirect burden: In burden of disease analysis, where the disease of interest is considered to be a risk factor (that is, disease-as-risk) for associated or ‘linked’ diseases, it is the burden from these linked diseases due to the disease-as-risk. For example, diabetes is considered to be a risk factor for coronary heart disease, stroke, dementia and other diseases, so the indirect burden is the burden attributable to diabetes for these linked diseases.
72 Diabetes and chronic kidney disease as risks for other diseases linked disease: Many risk factors are associated with developing certain diseases; for example, diabetes is associated with increased risk of developing coronary heart disease. The disease in association is a linked disease to the risk factor. meta-analysis: A statistical technique for combining findings from previous independent studies. It provides a quantitative estimate of the overall effect of an intervention or variable on a defined outcome, giving due weight to the size of the different studies included. morbidity: Ill health in an individual, and levels of ill health in a population or group. mortality: Death. non-fatal burden: The burden from living with ill health as measured by years lived with disability. Often used synonymously with YLD, and also referred to as ‘health loss’ in this report. odds ratio (OR): Odds ratio is a measure of association which compares the odds of disease in those exposed, to the odds of disease in those unexposed. population attributable fraction (PAF): The proportion (fraction) of a disease, illness, disability or death in a population that can be attributed to a particular risk factor or combination of risk factors. prevalence: The number of cases of a disease or injury in a population at a given time. relative risk (RR): The risk of an event relative to exposure, calculated as the ratio of the probability of the event occurring in the exposed group to the probability of it occurring in the non-exposed group. A relative risk of 1 implies no difference in risk; a RR <1 implies the event is less likely to occur in the exposed group; and a RR >1 implies the event is more likely to occur in the exposed group. risk factor (for health): Any factor that causes or increases the likelihood of a health disorder or other unwanted condition or event. risk–outcome pairs: Conditions that are causally linked to a risk factor. YLD (years lived with disability): A measure of the years of what could have been a healthy life but were instead spent in states of less than full health. YLD represent non-fatal burden. YLL (years of life lost): Years of life lost due to premature death, defined as dying before the global ideal life span. YLL represent fatal burden.
Diabetes and chronic kidney disease as risks for other diseases 73
References
ABS (Australian Bureau of Statistics) 2013. Population projections, Australia, 2012 (base) to 2101. ABS cat. no. 3222.0. Canberra: ABS. Afkarian M, Sachs MC, Kestenbaum B, Hirsch IB, Tuttle KR, Himmelfarb J et al. 2013. Kidney disease and increased mortality risk in type 2 diabetes. Journal of the American Society of Nephrology 24:302–308. AIHW (Australian Institute of Health and Welfare) 2011. Key indicators of progress for chronic disease and associated determinants: data report. Cat. no. PHE 142. Canberra: AIHW. AIHW 2012a. National Healthcare Agreement: Performance Indicator 03—incidence of end-stage kidney disease, 2012 Quality Statement. Viewed 10 July 2016,
74 Diabetes and chronic kidney disease as risks for other diseases
Baber U, Gutierrez OM, Levitan EB, Warnock DG, Farkouh ME, Tonelli M et al. 2013. Risk for recurrent coronary heart disease and all-cause mortality among individuals with chronic kidney disease compared with diabetes mellitus, metabolic syndrome and cigarette smokers. American Heart Journal 166(2):373–380. Baker LD, Cross DJ, Minoshima S, Belongia D, Watson GS & Craft S 2011. Insulin resistance and Alzheimer-like reductions in regional cerebral glucose metabolism for cognitively normal adults with prediabetes or early type 2 diabetes. Archives of Neurology 68:51–7. Bao C, Yang X, Xu W, Luo H, Xu Z, Su C et al. 2013. Diabetes mellitus and incidence and mortality of kidney cancer: a meta-analysis. Journal of Diabetes 27:357–64. Barnes DE & Yaffe K 2011. The projected effect of risk factor reduction on Alzheimer’s disease prevalence. The Lancet Neurology 10:819–28. Baronea M & Menna-Barretob L 2011. Diabetes and sleep: a complex cause-and-effect relationship. Diabetes Research and Clinical Practice 91(2):129–137. Bedse G, Di Domenico F, Serviddio G & Cassano T 2015. Aberrant insulin signaling in Alzheimer’s disease: current knowledge. Frontiers in Neuroscience 9:204. Biessels GJ, Staekenborg S, Brunner E, Brayne C & Scheltens P 2006. Risk of dementia in diabetes mellitus: a systematic review. The Lancet Neurology 5:64–74. Boyle P, Boniol M, Koechlin A, Robertson C, Valentini F, Coppens K et al. 2012. Diabetes and breast cancer risk: a meta-analysis. British Journal of Cancer 107:1608–17. Bugnicourt JM, Godefroy O, Chillon JM, Choukroun G & Massy ZA 2013. Cognitive disorders and dementia in CKD: the neglected kidney-brain axis. Journal of the American Society of Nephrology 24:353–63. Cheng D, Fei Y, Liu Y, Li J, Xue Q, Wang X et al. 2014. HbA1C variability and the risk of renal status progression in diabetes mellitus: a meta-analysis. Plos One 9:e115509. Cheng G, Huang C, Deng H & Wang H 2012. Diabetes as a risk factor for dementia and mild cognitive impairment: a meta-analysis of longitudinal studies. Internal Medicine Journal 42:484–91. Cukierman T, Gerstein HC & Williamson JD 2005. Cognitive decline and dementia in diabetes—systematic overview of prospective observational studies. Diabetologia 48:2460–9. Deckers K, van Boxtel MP, Schiepers OJ, de Vugt M, Munoz Sanchez JL, Anstey KJ et al. 2015. Target risk factors for dementia prevention: a systematic review and Delphi consensus study on the evidence from observational studies. International Journal of Geriatric Psychiatry 30:234–46. Deng L, Gui Z, Zhao L, Wang J & Shen L 2012. Diabetes mellitus and the incidence of colorectal cancer: an updated systematic review and meta-analysis. Digestive Diseases and Sciences 57:1576–85. El–Serag HB, Hampel H & Javadi F 2006. The association between diabetes and hepatocellular carcinoma: a systematic review of epidemiologic evidence. Clinical Gastroenterology and Hepatology 4:369–80. Emerging Risk Factors Collaboration 2010. Diabetes mellitus, fasting blood glucose concentration and risk of vascular disease: a collaborative meta-analysis of 102 prospective studies. The Lancet 375:2215–22.
Diabetes and chronic kidney disease as risks for other diseases 75
Etgen T 2015. Kidney disease as a determinant of cognitive decline and dementia. Alzheimers Research and Therapy 7:29. Fowkes FG, Rudan D, Rudan I, Aboyans V, Denenberg JO, McDermott MM et al. 2013. Comparison of global estimates of prevalence and risk factors for peripheral artery disease in 2000 and 2010: a systematic review and analysis. The Lancet 382:1329–40. Friberg E, Orsini N, Mantzoros C & Wolk A 2007. Diabetes mellitus and risk of endometrial cancer: a meta-analysis. Diabetologia 50:1365–74. Gansevoort RT, Correa-Rotter R, Hemmelgarn BR, Jafar TH, Heerspink HJ, Mann JF et al. 2013. Chronic kidney disease and cardiovascular risk: epidemiology, mechanisms, and prevention. The Lancet 382:339–52. GBD 2013 Risk Factors Collaborators 2015. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990–2013; 2013: a systematic analysis for the Global Burden of Disease Study 2013: Supplementary appendix. The Lancet 386(10010):2287–2323. doi: 10.1016/S0140-6736(15)60692–4. Gudala K, Bansal D, Schifano F & Bhansali A 2013. Diabetes mellitus and risk of dementia: a meta-analysis of prospective observational studies. Journal of Diabetes Investigation 4:640–50. Guraya SY 2015. Association of type 2 diabetes mellitus and the risk of colorectal cancer: a meta-analysis and systematic review. World Journal of Gastroenterology 21:6026. Helmer C, Stengel B, Metzger M, Froissart M, Massy ZA, Tzourio C et al. 2011. Chronic kidney disease, cognitive decline, and incident dementia: the 3C Study. Neurology 77:2043–51. Herzog CA, Asinger RW, Berger AK, Charytan DM, Díez J, Hart RG et al. 2011. Cardiovascular disease in chronic kidney disease. A clinical update from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney International 80:572–86. Hiteshi AK, Li D, Gao Y, Chen A, Flores F, Mao SS et al. 2014. Gender differences in coronary artery diameter are not related to body habitus or left ventricular mass. Clinical Cardiology 37(10):605–609. Huxley R, Ansary-Moghaddam A, De González AB, Barzi F & Woodward M 2005. Type-II diabetes and pancreatic cancer: a meta-analysis of 36 studies. British Journal of Cancer 92:2076–83. Janghorbani M, Hu FB, Willett WC, Li TY, E MJ, Logroscino G et al. 2007. Prospective study of type 1 and type 2 diabetes and risk of stroke subtypes. Diabetes Care 30:1730–5. Kadiyala R, Peter R & Okosieme OE 2010. Thyroid dysfunction in patients with diabetes: clinical implications and screening strategies. International Journal of Clinical Practice 64(8):1130–1139. Klausen K, Borch-Johnsen K, Feldt-Rasmussen B, Jensen G, Clausen P, Scharling H et al. 2004. Very low levels of microalbuminuria are associated with increased risk of coronary heart disease and death independently of renal function, hypertension, and diabetes. Circulation 110:32–35.
76 Diabetes and chronic kidney disease as risks for other diseases
Kokubo Y, Nakamura S, Okamura T, Yoshimasa Y, Makino H, Watanabe M et al. 2009. Relationship between blood pressure category and incidence of stroke and myocardial infarction in an urban Japanese population with and without chronic kidney disease: the Suita Study. Stroke 40:2674–9. Larsson S & Wolk A 2011. Diabetes mellitus and incidence of kidney cancer: a meta-analysis of cohort studies. Diabetologia 54:1013–8. Larsson S, Orsini N, Brismar K & Wolk A 2006. Diabetes mellitus and risk of bladder cancer: a meta-analysis. Diabetologia 49:2819–23. Larsson SC, Mantzoros CS & Wolk A 2007. Diabetes mellitus and risk of breast cancer: a meta‐analysis. International Journal of Cancer 121:856–62. Larsson SC, Orsini N & Wolk A 2005. Diabetes mellitus and risk of colorectal cancer: a meta-analysis. Journal of the National Cancer Institute 97:1679–87. Lee M, Saver JL, Chang K-H & Ovbiagele B 2010b. Level of albuminuria and risk of stroke: systematic review and meta-analysis. Cerebrovascular Diseases 30:464–9. Lee M, Saver JL, Chang KH, Liao HW, Chang SC & Ovbiagele B 2010a. Low glomerular filtration rate and risk of stroke: meta-analysis. The BMJ 341:c4249–58. Liao CC, Shih CC, Yeh CC, Chang YC, Hu CJ, Lin JG et al. 2015. Impact of diabetes on stroke risk and outcomes: two nationwide retrospective cohort studies. Medicine 94:e2282. Lo SF, Chang SN, Muo CH, Chen SY, Liao FY, Dee SW et al. 2013. Modest increase in risk of specific types of cancer types in type 2 diabetes mellitus patients. International Journal of Cancer 132:182–8. Lu FP, Lin KP & Kuo HK 2009. Diabetes and the risk of multi-system aging phenotypes: a systematic review and meta-analysis. PLoS One 4:e4144. Luchsinger JA 2010. Diabetes, related conditions, and dementia. Journal of Neurological Sciences 299:35–8. MacKenzie T, Zens MS, Ferrara A, Schned A & Karagas MR 2011. Diabetes and risk of bladder cancer. Cancer 117:1552–6. Mankovsky BN & Ziegler D 2004. Stroke in patients with diabetes mellitus. Diabetes Metabolism Research and Reviews 20:268–87. Mao Y, Tao M, Jia X, Xu H, Chen K, Tang H et al. 2015. Effect of diabetes mellitus on survival in patients with pancreatic cancer: a systematic review and meta-analysis. Scientific Reports 5:17102. Martín-Timón I, Sevillano-Collantes C, Segura-Galindo A & Cañizo-Gómez FJ 2014. Type 2 diabetes and cardiovascular disease: have all risk factors the same strength? World Journal of Diabetes 5(4):444–470. Masson P, Webster AC, Hong M, Turner R, Lindley RI, Craig JC 2015. Chronic kidney disease and the risk of stroke: a systematic review and meta-analysis. Nephrology Dialysis Transplantation 30:1162–9. Matsushita K, van der Velde M, Astor BC, Woodward M, Levey AS, de Jong PE et al. 2010. Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: a collaborative meta-analysis. The Lancet 375:2073–81.
Diabetes and chronic kidney disease as risks for other diseases 77
Miwa K, Tanaka M, Okazaki S, Furukado S, Yagita Y, Sakaguchi M et al. 2014. Chronic kidney disease is associated with dementia independent of cerebral small-vessel disease. Neurology 82:1051–7. MOH (New Zealand Ministry of Health) 2012. Ways and means: a report on methodology from the New Zealand Burden of Diseases, Injuries and Risk Factors Study, 2006–2016. Wellington: New Zealand Ministry of Health. Montemarano N, Guttman J & McFarlane SI 2013. Anemia of chronic kidney disease—a modifiable risk factor in a growing high cardiovascular risk population. In Masuo K (ed.) Type 2 Diabetes. Rijeka: InTech. Najarian RM, Sullivan LM, Kannel WB, Wilson PW, D’Agostino RB & Wolf PA 2006. Metabolic syndrome compared with type 2 diabetes mellitus as a risk factor for stroke: the Framingham Offspring Study. Archives of Internal Medicine 166:106–11. Ninomiya T 2014. Diabetes mellitus and dementia. Current Diabetes Reports 14:487. Ninomiya T, Perkovic V, Verdon C, Barzi F, Cass A, Gallagher M et al. 2009. Proteinuria and stroke: a meta-analysis of cohort studies. American Journal of Kidney Diseases 53:417–25. Norton S, Matthews FE, Barnes DE, Yaffe K & Brayne C 2014. Potential for primary prevention of Alzheimer’s disease: an analysis of population-based data. The Lancet Neurology 13:788–94. O’Hare AM, Glidden DV, Fox CS & Hsu C-Y 2004. High prevalence of peripheral arterial disease in persons with renal insufficiency results from the National Health and Nutrition Examination Survey 1999–2000. Circulation 109:320–3. Perkovic V, Verdon C, Ninomiya T, Barzi F, Cass A, Patel A et al. 2008. The relationship between proteinuria and coronary risk: a systematic review and meta-analysis. PLoS Medicine 5:e207. Peters SA, Huxley RR & Woodward M 2014a. Diabetes as risk factor for incident coronary heart disease in women compared with men: a systematic review and meta-analysis of 64 cohorts including 858,507 individuals and 28,203 coronary events. Diabetologia 57:1542–51. Peters SA, Huxley RR & Woodward M 2014b. Diabetes as a risk factor for stroke in women compared with men: a systematic review and meta-analysis of 64 cohorts, including 775,385 individuals and 12,539 strokes. The Lancet 383:1973–80. Prince M, Albanese E, Guerchet M & Prina M 2014. World alzheimer report 2014—dementia and risk reduction: an analysis of protective and modifiable factors. London: Alzheimer’s Disease International. Profenno LA, Porsteinsson AP & Faraone SV 2010. Meta-analysis of Alzheimer’s disease risk with obesity, diabetes, and related disorders. Biological Psychiatry 67:505–12. Roberts RO, Knopman DS, Cha RH, Mielke MM, Pankratz VS, Boeve BF et al. 2014. Diabetes and elevated hemoglobin A1c levels are associated with brain hypometabolism but not amyloid accumulation. Journal of Nuclear Medicine 55:759–64. Rönnemaa E, Zethelius B, Lannfelt L & Kilander L 2011. Vascular risk factors and dementia: 40-year follow-up of a population-based cohort. Dementia and Geriatric Cognitive Disorders 31:460–6.
78 Diabetes and chronic kidney disease as risks for other diseases
Ryoo JH, Kim SG, Suh BS, Kim DI, Park SK 2011. Relationship between chronic kidney disease and risk of coronary heart disease in Korean men. Journal of Korean Medical Science 26(6):753–758. Sarnak MJ, Levey AS, Schoolwerth AC, Coresh J, Culleton B, Hamm LL et al. 2003. Kidney disease as a risk factor for development of cardiovascular disease a statement from the American Heart Association Councils on kidney in cardiovascular disease, high blood pressure research, clinical cardiology, and epidemiology and prevention. Circulation 108:2154–69. Sasaki Y, Marioni R, Kasai M, Ishii H, Yamaguchi S & Meguro K 2011. Chronic kidney disease: a risk factor for dementia onset: a population-based study. The Osaki-Tajiri Project. Journal of the American Geriatrics Society 59:1175–81. Savva GM & Stephan BC 2010. Epidemiological studies of the effect of stroke on incident dementia: a systematic review. Stroke 41:e41–6. Schiffrin EL, Lipman ML, Mann JF 2007. Chronic kidney disease: effects on the cardiovascular system. Circulation 116:85–97. Seliger SL, Siscovick DS, Stehman-Breen CO, Gillen DL, Fitzpatrick A, Bleyer A et al. 2004. Moderate renal impairment and risk of dementia among older adults: the Cardiovascular Health Cognition Study. Journal of the American Society of Nephrology 15:1904–11. Selvin E, Coresh J, Golden SH, Brancati FL, Folsom AR & Steffes MW 2005. Glycemic control and coronary heart disease risk in persons with and without diabetes—the Atherosclerosis Risk in Communities Study. Archives of Internal Medicine 165:1910–6. Shou J, Zhou L, Zhu S & Zhang X 2015. Diabetes is an independent risk factor for stroke recurrence in stroke patients: a meta-analysis. Journal of Stroke and Cerebrovascular Diseases 24:1961–8. Soedamah-Muthu SS, Fuller JH, Mulnier HE, Raleigh VS, Lawrenson RA & Colhoun HM 2006. High risk of cardiovascular disease in patients with type 1 diabetes in the UK. A cohort study using the general practice research database. Diabetes Care 29:798–804. Song S, Wang B, Zhang X, Hao L, Hu X, Li Z et al. 2015. Long-term diabetes mellitus is associated with an increased risk of pancreatic cancer: a meta-analysis. Plos One 10:e0134321. Stengel B 2010. Chronic kidney disease and cancer: a troubling connection. Journal of Nephrology 23:253–62. Suh S & Kim K-W 2011. Diabetes and cancer: is diabetes causally related to cancer? Diabetes and Metabolism Journal 35:193–8. Tonelli M, Muntner P, Lloyd A, Manns BJ, Klarenbach S, Pannu N et al. 2012. Risk of coronary events in people with chronic kidney disease compared with those with diabetes: a population-level cohort study. The Lancet 380:807–14. Tong B & Stevenson C 2007. Comorbidity of cardiovascular disease, diabetes and chronic kidney disease in Australia. Cardiovascular disease series no. 28. Cat. no. CVD 37.Canberra: AIHW. Tuttolomondo A, Maida C, Maugeri R, Iacopino G & Pinto A 2015. Relationship between diabetes and ischemic stroke: analysis of diabetes-related risk factors for stroke and of specific patterns of stroke associated with diabetes mellitus. Journal of Diabetes and Metabolism 6:544–50.
Diabetes and chronic kidney disease as risks for other diseases 79
Van der Velde M, Matsushita K, Coresh J, Astor BC, Woodward M, Levey A et al. 2011. Lower estimated glomerular filtration rate and higher albuminuria are associated with all-cause and cardiovascular mortality. A collaborative meta-analysis of high-risk population cohorts. Kidney International 79:1341–52. Vigneri P, Frasca F, Sciacca L, Pandini G & Vigneri R 2009. Diabetes and cancer. Endocrine-Related Cancer 16:1103–23. Vulesevic B, Milne RW & Suuronen EJ 2014. Reducing methylglyoxal as a therapeutic target for diabetic heart disease. Biochemical Society Transactions 42(2):523–527. Wang F, Herrington M, Larsson J & Permert J 2003. The relationship between diabetes and pancreatic cancer. Molecular Cancer 2:4–8. Wang P, Kang D, Cao W, Wang Y & Liu Z 2012. Diabetes mellitus and risk of hepatocellular carcinoma: a systematic review and meta‐analysis. Diabetes/Metabolism Research and Reviews 28:109–22. Wang YG, Wang P, Wang B, Fu ZJ, Zhao WJ & Yan SL 2014. Diabetes mellitus and poorer prognosis in hepatocellular carcinoma: a systematic review and meta-analysis. PLoS One 9:e95485. WCRF (World Cancer Research Fund) & AICR (American Institute of Cancer) 2007. Food, nutrition, physical activity and the prevention of cancer: a global perspective. Washington: AICR. Weiner DE, Tighiouart H, Amin MG, Stark PC, Macleod B, Griffith JL et al. 2004. Chronic kidney disease as a risk factor for cardiovascular disease and all-cause mortality: A pooled analysis of community-based studies. Journal of the American Society of Nephrology 15:1307–1315. WHO (World Health Organization) 2013. Global Action Plan for the Prevention and Control of Non-communicable Diseases 2013–2020. Geneva: WHO. Xu X, Wu J, Mao Y, Zhu Y, Hu Z, Lin Y et al. 2013. Diabetes mellitus and risk of bladder cancer: a meta-analysis of cohort studies. PLoS One 8:e58079. Yang WS, Va P, Bray F, Gao S, Gao J, Li HL et al. 2011. The role of pre-existing diabetes mellitus on hepatocellular carcinoma occurrence and prognosis: a meta-analysis of prospective cohort studies. Plos One 6:e27326. Yang YX, Hennessy S & Lewis JD 2004. Insulin therapy and colorectal cancer risk among type 2 diabetes mellitus patients. Gastroenterology 127:1044–50. Zhu Z, Wang X, Shen Z, Lu Y, Zhong S & Xu C 2013. Risk of bladder cancer in patients with diabetes mellitus: an updated meta-analysis of 36 observational studies. BioMed Central Cancer 13:310–17.
80 Diabetes and chronic kidney disease as risks for other diseases
List of tables
Table 2.1: Risk factor population exposure definition, 2011 ...... 11 Table 3.1: Diabetes direct burden, by sex, 2011 ...... 14 Table 3.2: Collective burden (DALY) of diabetes, by direct and indirect burden and sex, 2011 ...... 15 Table 3.3: Attributable burden due to diabetes, by linked disease and sex, 2011 ...... 17 Table 3.4: Estimated diabetes burden in 2020 under different exposure scenarios ...... 29 Table 4.1: Chronic kidney disease direct burden, by sex, 2011 ...... 31 Table 4.2: Collective burden (DALY) of chronic kidney disease, by direct and indirect burden and sex, 2011 ...... 32 Table 4.3: Attributable burden due to chronic kidney disease, by linked disease and sex, 2011 ...... 33 Table 4.4: Estimated chronic kidney disease burden in 2020 under different exposure scenarios ...... 40 Table 5.1: Comparison of direct and indirect diabetes burden between Australia and New Zealand diseases-as-risks analyses ...... 43 Table A1: Effect size and sources for diabetes and linked disease pair analysis...... 47 Table A2: Source of effect size used in calculation of attributable burden for diabetes ...... 53 Table B1: Effect size and sources for chronic kidney disease and linked disease pair analysis...... 56 Table B2: Source of effect size used in calculation of attributable burden for chronic kidney disease ...... 59 Table C1: Estimated diabetes prevalence rates (number per 1,000) for 2003 and 2011 used in scenario modelling ...... 63 Table C2: Diabetes YLL rates (number per 1,000) for 2003 and 2011 used in scenario modelling ...... 64 Table C3: Estimated chronic kidney disease prevalence rates (number per 1,000) for 2003 and 2011 used in scenario modelling ...... 65 Table C4: Chronic kidney disease YLL rates (number per 1,000) for 2003 and 2011 used in scenario modelling ...... 66 Table C5: Indirect diabetes burden estimated in 2020 under different prevalence scenarios, by linked disease ...... 67 Table C6: Indirect chronic kidney disease burden under different exposure scenarios, by linked disease ...... 68 Table D1: The Australian Burden of Disease Study 2011 main data sources for YLD estimation ...... 71
Diabetes and chronic kidney disease as risks for other diseases 81
List of figures
Figure 1.1: Examples of direct and indirect burden of diabetes ...... 1 Figure 1.2: An example of possible causal pathways of diabetes and CKD to associated diseases ...... 4 Figure 3.1: Diabetes direct burden, by age and sex, 2011 ...... 14 Figure 3.2: Collective burden (DALY) of diabetes, by direct and indirect burden, by age and sex, 2011 ...... 16 Figure 3.3: Proportion of burden attributable to diabetes, by linked disease and sex, 2011 ...... 18 Figure 3.4: Proportion of indirect diabetes burden, by linked disease due to fatal and non-fatal outcomes, 2011 ...... 19 Figure 3.5: Proportion of indirect diabetes burden due to fatal and non-fatal outcomes, by age, 2011 ...... 19 Figure 3.6: Burden (DALY) of coronary heart disease attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 20 Figure 3.7: Burden (DALY) of stroke attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 21 Figure 3.8: Burden (DALY) of peripheral vascular disease attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 22 Figure 3.9: Burden (DALY) of liver cancer attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 23 Figure 3.10: Burden (DALY) of pancreatic cancer attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 23 Figure 3.11: Burden (DALY) of bowel cancer attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 24 Figure 3.12: Burden (DALY) of breast and uterine cancers attributable to diabetes, by (a) number and (b) proportion of burden, by age and cancer type, 2011 ...... 25 Figure 3.13: Burden (DALY) of kidney and bladder cancer attributable to diabetes, by (a) number and (b) proportion of burden, by age and cancer type, 2011 ...... 26 Figure 3.14: Burden (DALY) of dementia attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 27 Figure 3.15: Burden (DALY) of chronic kidney disease attributable to diabetes, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 27 Figure 3.16: Estimated indirect diabetes burden (DALY) in 2020 under different exposure scenarios, by age ...... 30 Figure 4.1: Chronic kidney disease direct burden, by age and sex, 2011 ...... 31 Figure 4.2: Collective burden (DALY) of chronic kidney disease, by direct and indirect burden, by age and sex, 2011 ...... 33 Figure 4.3: Proportion of burden attributable to chronic kidney disease, by linked disease and sex, 2011 ...... 34 Figure 4.4: Proportion of indirect chronic kidney disease burden, by linked disease due to fatal and non-fatal outcomes, 2011 ...... 35
82 Diabetes and chronic kidney disease as risks for other diseases
Figure 4.5: Proportion of indirect chronic kidney disease burden due to fatal and non-fatal outcomes, by age, 2011 ...... 35 Figure 4.6: Burden (DALY) of coronary heart disease attributable to chronic kidney disease, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 36 Figure 4.7: Burden (DALY) of stroke attributable to chronic kidney disease, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 37 Figure 4.8: Burden (DALY) of dementia attributable to chronic kidney disease, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 38 Figure 4.9: Burden (DALY) of peripheral vascular disease attributable to chronic kidney disease, by (a) number and (b) proportion of burden, by age and sex, 2011 ...... 38 Figure 4.10: Estimated indirect chronic kidney disease burden (DALY) in 2020 under different exposure scenarios, by age ...... 41
Diabetes and chronic kidney disease as risks for other diseases 83
List of boxes
Box 1.1: Diabetes and chronic kidney disease ...... 3 Box 1.2: Key terms used in this report ...... 6 Box 2.1: World Cancer Research Fund criteria for level of evidence ...... 9 Box 2.2: Example of how a population attributable fraction is applied to the population ...... 12
84 Diabetes and chronic kidney disease as risks for other diseases
Related publications
This report, Diabetes and chronic kidney disease as risks for other diseases, and other AIHW publications can be downloaded for free from the AIHW website
Diabetes and chronic kidney disease as risks for other diseases 85 This report aims to provide a more comprehensive picture of the full health loss attributable to diabetes and chronic Diabetes and chronic kidney disease kidney disease (CKD). It quantifies the impact of diabetes and CKD on the burden of other diseases for which there as risks for other diseases is evidence of a causal association (‘linked diseases’) to estimate the indirect burden caused by these 2 diseases. It uses disease burden estimates from the Australian Burden of Disease Study 2011 and extends the standard Australian Burden of Disease Study 2011 approach for analysis of risk factors to model diabetes and CKD as risk factors. When the indirect burden due to linked diseases was taken into account, the collective burden due to diabetes was 1.9 times as high, and CKD was 2.1 times as high, as their direct burden.
AUSTRALIAN BURDEN OF DISEASE STUDY SERIES NO. 8